The California Electricity Crisis (HOOVER INST PRESS PUBLICATION)

THE CALIFORNIA ELECTRICITY CRISIS: WHAT, WHY, AND WHAT’S NEXT This page intentionally left blank THE CALIFORNIA ELE...

Author: James L. Sweeney

19 downloads 685 Views 6MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

THE CALIFORNIA ELECTRICITY CRISIS: WHAT, WHY, AND WHAT’S NEXT

This page intentionally left blank

THE CALIFORNIA ELECTRICITY CRISIS: WHAT, WHY, AND WHAT’S NEXT

Charles J. Cicchetti School of Policy, Planning, and Development University of Southern California Los Angeles, California

Jeffrey A. Dubin Division of Humanitiies and Social Sciences California Institute of Technology Pasadena, California

Colin M. Long Pacific Economics Group, L.L.C.

KLUWER ACADEMIC PUBLISHERS NEW YORK, BOSTON, DORDRECHT, LONDON, MOSCOW

eBook ISBN: Print ISBN:

1-4020-8032-8 1-4020-7692-4

©2004 Springer Science + Business Media, Inc. Print ©2004 Kluwer Academic Publishers Boston All rights reserved No part of this eBook may be reproduced or transmitted in any form or by any means, electronic, mechanical, recording, or otherwise, without written consent from the Publisher Created in the United States of America

Visit Springer's eBookstore at: and the Springer Global Website Online at:

http://www.ebooks.kluweronline.com http://www.springeronline.com

Dedication

This book is dedicated to Sally, Jackie, and Molly.

This page intentionally left blank

Contents

List of Tables and Figures

ix

List of Terms

xiii

Preface

xvii

1. Introduction

1

2. Power Production Economics

5

3. Principles of Traditional Regulation

15

4. Reconciling Marginal Cost and Revenue Requirements

25

5. Competitive Wholesale Markets for Electricity

47

6. California’s Market Design: an Initial Success Followed by a “Perfect Storm”

53

7. Design Flaws and a Worsening Crisis

63

viii

The California Electricity Crisis: What, Why, and What’s Next

8. Testable Hypothesis

71

9. Survey of Electricity Models for California

75

10. An Economic Analysis of Natural Gas Price Movements During the Crisis

85

11. An Econometric Analysis of Electricity Prices in California

103

12. Market Manipulation

127

13. Gaming and Cheating

137

14. Market Monitoring and Initial Regulatory Response

147

15. Refunds and Mitigation

159

16. California Responds

173

17. Handicapping the Winners

185

18. Conclusion: Wrapping Up and Lessons Learned

195

Index

203

About the Authors

207

List of Tables and Figures

Tables Table 5-1.

Economic Actions of Sellers

Table 5-2.

Economic Actions of Buyers

Table 5-3.

Economic Choices of the Power Exchange

Table 5-4.

Transactions that Result in this Competitive Market

Table 7-1.

Market Hedges Compared to the Spot Market in Other Deregulated Electricity Markets

Table 7-2.

Volume of Megawatts Purchased Out of Market, June – December 2000

Table 9-1.

Quantitative Analyses Relevant to the California Energy Crisis

Table 10-1.

Interstate Deliveries to California

Table 10-2.

California Interstate Pipeline System

Table 10-3.

List of Variables

Table 10-4.

Daily Natural Gas Price Regressions

Table 10-5.

Daily Natural Gas Price Regressions for Simulations

Table 10-6.

Average Monthly Actual and Predicted Southern California Natural Gas Prices

Table 10-7.

Average Monthly Actual and Predicted Northern California Natural Gas Prices

x

The California Electricity Crisis: What, Why, and What’s Next

Table 11-1.

Explanatory Variables for Daily Electricity Equations

Table 11-2.

Dependent Variables and Corresponding Descriptions

Table 11-3.

Logit Model for Endogenous Event Day Factor

Table 11-4.

Regression Models for Daily Electricity Prices in California Markets

Table 11-5.

Regression Models for Daily Electricity Prices in Western Markets

Table 14-1.

Potential Concerns Identified by the Market-Monitoring Groups

Table 15-1.

Approximate Amounts Owed and Owing

Table 15-2.

Weekly Average Comparison of Natural Gas Prices Established by the CAISO and the FERC Staff for Northern and Southern California

Table 15-3.

Weekly Average Comparison of the CAISO’s October 2001 MMCP and the FERC Staff’s MMCP

Table 16-1.

Projected SCAG Region Electricity Use

Figures Figure 2-1.

Representative Hourly Load (Descending

Figure 2-2.

Levels of Hourly Load

Figure 2-3.

Representative Costs of Various Plant Types

Figure 4-1.

Supply and Demand in a Competitive Marketplace

Figure 4-2.

Firm Cost Functions

Figure 4-3.

Price Levels Influencing Firm Behavior

Figure 4-4.

Long-Run Average Cost Function Based on Various Levels of Firm Investment

Figure 4-5.

Supply and Demand of a Monopolistic Firm

Figure 4-6.

Decreasing Cost Case for the ith Consumer Group

The California Electricity Crisis: What, Why, and What’s Next Figure 5-1.

Physical Flows in an ISO/PX Market

Figure 5-2.

Merit Order of Five Bids to Sell KWs

Figure 5-3.

Willingness to Pay

Figure 5-4.

Stylized Market Clearing

Figure 6-1.

Comparison of Growth in Peak Demand Versus Growth in Generating Capacity, 1996-2000

Figure 6-2.

January to July Volumes of Run-Off in the Northwest, 1992-2000

Figure 6-3.

CAISO Actual Average Demand from May to August, 1998-2000

Figure 6-4.

Natural Gas Spot Price* ($/Mcf), Mar 1998 – Dec 2000

Figure 9-1.

Short-Run Marginal Cost

Figure 10-1.

Southern California Natural Gas Prices

Figure 10-2.

Northern California Natural Gas Prices

Figure 10-3.

Southern California Natural Gas Prices (Jan 2000 - Jun 2001)

Figure 10-4.

Northern California Natural Gas Prices (Jan 2000 - Jun 2001)

Figure 11-1.

SP-15 Peak Prices

Figure 11-2.

SP-15 Off-Peak Prices

Figure 11-3.

SP-15 Peak Prices

Figure 11-4.

NP-15 Off-Peak Prices

Figure 11-5.

PX Peak Prices

Figure 11-6.

PX Off-Peak Prices

Figure 11-7.

Mid-Columbia Peak Prices

Figure 11-8.

Mid-Columbia Off-Peak Prices

Figure 11-9.

California-Oregon Border Peak Prices

Figure 11-10.

California-Oregon Border Off-Peak Prices

Figure 11-11.

Four Corners Peak Prices

Figure 11-12.

Four Corners Off-Peak Prices

Figure 11-13.

Palo Verde Peak Prices

Figure 11-14.

Palo Verde Off-Peak Prices

xi

This page intentionally left blank

List of Terms

AB: AC: AFC: ALJ: ATC: AVC:

Assembly Bill Average Costs Average Fixed Costs Administrative Law Judge Average Total Cost Average Variable Cost

BCF: BGP:

Billion Cubic Feet Burbank, Glendale, Pasadena Power Utilities

CAISO: CAPM: CC: CCA: CCCT: CDWR: CEC: CEC Report: CF: COB: COS: CPA: CPUC: CPX : CRS :

California Independent System Operator Capital Asset Pricing Model Total Construction Cost (Contract Cost) Community Choice Aggregator Combined Cycle Combustion Turbine California Department of Water Resources California Energy Commission Natural Gas Infrastructure Issues Capacity Factor California-Oregon Border Cost of Service California Power Authority California Public Utility Commission California Power Exchange Cost Responsibility Surcharge

xiv

The California Electricity Crisis: What, Why, and What’s Next

D: DA: DCF: DISCO: DMA: DSM:

Outstanding Debt Direct Access Discounted Cash Flow Distribution Company Department of Market Analysis Demand Side Management

E: EPA: ESP:

Value of Equity Energy Policy Act of 1992 Electric Service Provider

FC: FERC: FPA:

Fixed Cost Federal Energy Regulatory Commission Federal Power Act

GLS:

Generalized Least Squares

HHI:

Hirschman-Herfindahl Index

IOU: IPP: ISO:

Investor-owned Utility Independent Power Producer Independent System Operator

JPA:

Joint Powers Authority

kW: kWH:

Kilowatt Kilowatt hour

LADWP: LOLP: LRAC:

Los Angeles Department of Water & Power Loss of Load Probability Long Run Average Cost

MAF: MC: MCC: MCF: MCP: Mid-C: MMBtu: MMCP: MMIP: MOU:

Million Acre Feet Marginal Cost Marginal Capacity Cost Million Cubic Feet Market Clearing Price Mid-Columbia Million British Thermal Units Mitigated Market Clearing Price Market Monitoring and Information Protocols Municipally Owned Utility

The California Electricity Crisis: What, Why, and What’s Next

xv

MPA: MTV: MW: MWH:

Municipal Power Authority MountainView Megawatt Megawatt Hour

NOx: NUG:

Nitrogen Oxide Non Utility Generator

O&M: OE: OLS: OOM:

Operation and Maintenance Annual Operating Expenses Ordinary Least Squares Out of Market

P: PG&E: PJM: PPA: PURPA: PV(F):

Price Pacific Gas & Electric Pennsylvania, New Jersey, Maryland Purchase Power Agreement Public Utility Regulatory Policy Act of 1978 Present Value of Any Fuel Savings Due to Changes in the Merit Order of existing plants

Q: QF:

Quantity Sold Qualifying Facility

r: R: RB: ROE: ROR: RR:

Weighted average interest on debt Allowed regulatory profit Rate Base Return on Equity Rate of Return Revenue Requirement

SC: SCAG: SCE: SDG&E: SMD: SRAC: SRMC:

Scheduling Coordinator Southern California Association of Governments Southern California Edison San Diego Gas & Electric Standard Market Design Short Run Average Cost Short Run Marginal Cost

t: TC: TOU:

Corporate Tax Rate Annual Total Cost Time of Use

xvi

The California Electricity Crisis: What, Why, and What’s Next

TR:

Total Revenue

UMCP:

Unconstrained Market Clearing Price

VAR:

Vector Auto-Regression

WSCC:

Western States Coordinating Council

Preface

This book attempts to explain what went wrong in California’s restructured energy markets and what must be done to restore California’s economy and build new electricity systems. The intention here is to reconcile the principles of competition and regulation. California had a severe electricity crisis for about thirteen months beginning in May of 2000. The economic consequences and political fallout that arose from this crisis persist. California’s economy continues to suffer and the state’s treasury is deeply in debt. The state’s three investor-owned utilities were nearly financially decimated. San Diego Gas & Electric has recovered to a greater degree than the other two only because its retail prices are about three times the national average and, for a time, well above the other two IOUs in California. Southern California Edison has recently been restored to investment grade and was granted a rate increase. Pacific Gas & Electric is emerging from bankruptcy. This book discusses all of this in greater detail. The problems and consequences arising from California’s ill-fated foray into electricity market restructuring could damage the state for years to come. Challenges of this nature are not new to the Golden State. In the past, as we explain here, pragmatic, not entrenched, approaches have worked best in California. If California is to relatively quickly restore its previous enviable economic vitality and recover from the damage done to tarnish its luster, pragmatic approaches must again be used. We acknowledge the help of Marilea Fried at Kluwer and the useful comments of the referees. We especially thank Dan Mazmanian for urging us to write this book. This effort has benefited from past collaboration with Paul Joskow, Kerry Smith, William Gillen, Paul Smolensky, John Jurewitz,

xviii

The California Electricity Crisis: What, Why, and What’s Next

Irwin Stelzer, Bill Hogan, and Ralph Turvey. This work has also benefited from others who have worked on the problems in California, especially Chip Wright, Jon Hockenyos, Lawrence Acker, Cheryl Feik-Ryan, Cheryl Foley, Gordon Erspamer, Gary Bachman, Dan Flanagan, Jeff Kinsell, Len Viejo, and Howard Golub. Additionally, Art Lewbel provided helpful comments on the econometric sections along with anonymous reviewers. James Lin and Nicole Brackett provided excellent research and editorial assistance. The analysis and discussion presented here draws on many person years of research, analyses, and direct participation in much of the regulatory and legal matters discussed. These efforts include: working on and drafting the California State Audit Report that identified the various market and design problems that caused California’s energy crisis; analysis and expert testimony related to the California Refund and related legal proceedings; analysis of alleged gaming practices and testimony before the U.S. Congress; serving on Governor Davis’s Market Advisory Board; participating in the CPUC proceeding with respect to implementing AB 117; and preparing a White Paper and conducting two public policy forums on how California can fix some of these nagging market/regulatory problems. This book goes beyond our previous research and analysis and attempts to integrate what we have learned and to draw on the research of others. Any defects or omissions are our responsibility.

Chapter 1 INTRODUCTION

California is the world’s fifth largest economy. Its average and median percapita incomes are well above those in the rest of the states in the United States and all but four other countries in the rest of the world. The state is a magnet for entrepreneurs and innovation. Available energy is taken as a given. The state is rich in human capital. It benefits from a creative mix of cultures and religions from virtually every part of the planet. This diversity, plus ambition and a culture that rewards new ideas and products, makes California special. Life in the state is generally excellent in many important respects. The climate is particularly good–generally warm with relatively little humidity. The state’s ports move much of the world’s commerce. California sets worldwide trends in fashion, entertainment, technology, services, and many other consumer goods. Against this backdrop of good fortune and great economic success, the state experienced a dramatic energy crisis in 2000 and 2001, a crisis that by any standards was huge. This crisis was mostly unprecedented in California’s history, and was even more surprising when compared to California’s historic “golden touch” and youthful “can do” attitude. The state’s recent energy crisis, particularly in the electric sector, and the economic problems that followed have hit California hard. This circumstance was so unique that it helped propel an unprecedented recall of recently elected Governor Gray Davis, who was then sacked by the state’s voters. The state’s treasury is deeply in the red. The three large investor-owned utilities (IOUs) were pushed close to financial ruin. The state’s collective confidence has been shaken deeply. Nevertheless, people still flock to California, mostly from less fortunate places. These new immigrants want to work and they need housing, energy, transportation, education, and health

2

The California Electricity Crisis: What, Why, and What’s Next

services. These new challenges, coupled with the state’s recent energy and financial crisis, put many of the state’s diverse aspirations in jeopardy. This book explains “what” California was trying to do in the 1990s in order to reduce its energy costs. The details are critical because there are important lessons to be learned and steps to be avoided in other markets and other places. While most know that the California economic powerhouse was crippled by an electricity crisis in 2000, fewer realize that California’s complex energy industry and institutional restructuring was, by all accounts, deemed widely successful in 1998 and 1999. To avoid making similar mistakes in the future, it is important to understand what was done, why it was done, and review the program’s initial success. Consequently, this book addresses “why” things went so bad, so fast. 1 We start this discussion in Chapter 2 by undertaking a rudimentary review of the economics that drive power production. The starting point for all electricity restructuring is traditional comprehensive cost-of-service regulation. Thus, Chapter 3 discusses the underlying principles of traditional regulation. Chapter 4 explains how regulators have attempted to reconcile the economic principles associated with competitive markets and cost-of-service regulation principles. In Chapter 5, we discuss tariff reforms that are based on this reconciliation. Here, we explain how these reforms led to experiments in complex restructuring where competitive wholesale generation replaced comprehensive cost-of-service regulation. Chapter 6 uses California as an example to explain the basic features of this restructuring approach. Next, in Chapter 7, we turn to California’s specific market design, and explain how California’s restructuring went from a “good” start to a “bad” patch and ultimately culminated in an “ugly” conspiracy of mistakes, missteps, and bad luck. This “ugly” trifecta has been dubbed “The Perfect

1

Our approach is complimentary to other studies that have appeared in the literature, including Faruqui, et al. (2001), Joskow (2001), Borenstein (2002), Cicchetti, et al. (2001); Ahmad Faruqui, Hung-po Chao, Vic Niemeyer, Jeremy Platt, and Karl Stahlkopf, “Analyzing California’s Power Crisis,” The Energy Journal, Vol. 22, No. 4, 2001, pp. 2952; Paul Joskow, “California’s Electricity Crisis,” Oxford Review of Economic Policy, Vol. 17, No. 3, 2001, pp. 365-388; Severin Borenstein, “The Trouble with Electricity Markets: Understanding California’s Restructuring Disaster,” Journal of Economic Perspectives, Vol. 16. No. 1, 2002, pp. 191-211; Charles J. Cicchetti, Jeffrey A. Dubin, Jon Hockenyos, Colin M. Long, and J.A. Wright, “Energy Deregulation: The Benefits of Competition Were Undermined by Structural Flaws in the Market. Unsuccessful Oversight, and Uncontrollable Competitive Forces,” California State Auditor, Bureau of State Audits, Sacramento, California, March 2001.

Introduction

3

Storm” because virtually all the things that could have gone wrong did go wrong, seemingly at the same time from May 2000 through June 2001. Some of the worst problems were related to “design flaws and regulatory missteps.” These problems were exacerbated by climate and market forces. Chapter 8 takes a statistical approach to analyze what happened in California. In this chapter, several statistical hypotheses are stated. We then explain how we performed the econometric analyses. Chapter 9 reviews the modeling and approaches several others used to explain electricity prices in California. These efforts informed and influenced much of what we did. Chapter 10 analyzes natural gas, which is the primary marginal fuel used to generate electricity in California, and natural gas prices. In late 2000, natural gas prices snapped a decade and a half price slump, jumping spectacularly (five-fold) in North America, and particularly so (nearly thirtyfold) in California. We identify potential causes of these price spikes and then use an econometric analysis to test various hypotheses related to natural gas prices. Chapter 11 uses a similar econometric approach to analyze structural, regulatory, climatic, natural gas, and other economic factors that contributed to or caused electricity prices to surge from an average of about $30 per MWH in 1999 to $150 per MWH from mid-year 2000 to mid-year 2001. Potential causes for this price surge are identified and hypotheses tested in this chapter. To complete an analysis of “what and why,” we must go beyond reviewing the events leading up to and during the crisis. Thus, Chapter 12 discusses the role that market participants played in attempting to manipulate the market. Much has been said about gaming or attempting to game the market. These issues are reviewed in some detail in Chapter 13. Chapter 14 analyzes what the two primary regulatory bodies did during the crisis to bring the market under control. The California Public Utility Commission (CPUC) and the Federal Energy Regulatory Commission (FERC) were the key governmental entities charged with overseeing and controlling these markets. These regulatory agencies often were in conflict and, by all accounts, were slow to respond. This is an important sidebar within the main story of the electricity crisis, particularly because there are, ironically, new calls to expand regulation in response to what arguably was, at least partially, a major political/regulatory failure. Chapter 15 is a very important addendum to the regulatory actions taken during the crisis where we examine federal and state agencies’ efforts to revisit the “winners and losers” and recreate the prices that would have attained in a competitive market. In attempting to award refunds and grant

4

The California Electricity Crisis: What, Why, and What’s Next

some relief in an effort to neutralize the worst of the market’s fury, new “winners and losers” will be created. We examine these regulatory efforts in this chapter. Chapter 16 explains how the crisis spawned by the failures in 2000 continues to plague the state and threaten its ability to recover and to build the infrastructure needed for the state to grow. Here, we discuss how California is responding to these new challenges and boldly propose some new pragmatic hybrid solutions to achieve growth and regain prosperity Chapter 17 discusses how events in California are conspiring to handicap the winners. Chapter 18 details the lessons learned so as to guide others in their various pursuits to liberalize their energy industries.

Chapter 2 POWER PRODUCTION ECONOMICS

Historically, governments have determined that electricity should not be provided by competitive firms. Similar decisions have been made for natural gas, water, sewer, telephone, and cable service. These industries share several common characteristics. These are: A perception that the service is more or less essential and that the service coverage should be universal; Duplicative or parallel delivery systems might be too costly; Economies of scale are associated with a distribution system so that installation costs increase linearly with the radius of the wires or pipes while the volume that can be transported increases by the radius squared; Other high fixed costs, which may share the duplication and economy of scale characteristics, might also be central to service provision (e.g., power stations, natural gas storage, telephone switching centers, sewer treatment plants, etc.). Governments generally either decided to own the firms that have these characteristics or to permit private ownership in particular geographic areas, or franchises, subject to government regulation of price, profit, investment, and service quality. Government ownership of utilities was the primary choice outside the United States, while government regulation of privately owned utilities was more typical in the United States. Increasingly, the basic assumptions that led governments to rely on exclusive monopoly utility franchises began to change in the 1980s. Governmentally owned monopolies have been privatized. Competition has replaced the traditional cost-of-service approach for utility services. Customers are being given a choice of suppliers. Parts of the supply systems are being unbundled and made competitive. New competitive entrants are

6

The California Electricity Crisis: What, Why, and What’s Next

being encouraged at both retail and wholesale levels. Access to wires and pipes, which retain favorable single-supplier cost characteristics, is being opened to competing suppliers (electric generators, long distance telephone providers, natural gas producers) on an open access, comparable service terms, and non-discriminatory price basis. Much of the world recognizes that restructuring these basic, essential, and universal industries means that transition costs and other competitive issues must be resolved. The initial step is to assess the status quo, where public utilities or enterprises were granted a monopoly franchise. Pricing public enterprise or utility services is unusual because the monopoly franchise is not a price-taker. Instead, governments establish prices. The principal difference between utilities and firms that operate in competitive markets is how prices are established. Public utility prices have typically been set equal to their costs of service (COS), including some political notion of a fair return on the capital invested by private firms or a political assessment of a financial transfer or payment to the treasury. Public utilities minimize the costs associated with satisfying demand. This chapter explains the rudimentary aspects of the economics of power systems.

ECONOMIC PRINCIPLES AND POWER SYSTEMS Historically, governments controlled or owned power systems. Electricity suppliers were organized as virtually integrated systems that provided service to electric customers in a franchise territory without any retail competition. Typically, a single supplier owned all the power stations, transformers, power lines, dispatch or coordination centers, and distribution and metering equipment in a specific geographic or franchise area. Nevertheless, the same economic principles that guide a competitive firm’s quest for both operational and long-run investment efficiency apply equally to monopoly-owned, integrated power systems. The primary difference is that markets rely on demand and supply to set prices, while governments set utility prices based on the COS. Typically, governments ensure that the utility enterprise covers, but does not exceed, the fixed costs necessary to meet customer demand plus operating expenses. The term “covers” means both a return “of” investment and a fair return “on” that investment. Least-cost principles shape and determine both short-run operational efficiencies and long-run investment planning. The two aspects of electricity supply are energy and power. Energy is a flow concept that is

Power Production Economics

7

measured in kilowatt-hours (kWH) or megawatt hours (MWH). Power is a capacity concept that is measured in kWs or MWs at an instant in time. Power systems are built and operated to satisfy the demand for electricity, which varies by time of day, day of the week, and season of the year. Figure 2-1 depicts a representative power system’s demand or load, ranging from the highest hourly demand at A to the lowest hourly demand at C, over the course of a year.

Specifically, Figure 2-1 shows the hours in a year where demand is at or above the MWs shown on the vertical axis. At point A, demand (load) is at its peak. This may happen over a very few hours in the year. At point B, demand (load) is about the average demand level for the year. B' is the median demand for the year. In other words, B' is the demand level that corresponds to the midpoint in the year, or 8760 / 2 = 4380 hours. B' represents the point at which half the year’s electricity demand is greater than B' and half the year’s electricity demand is less. Finally, point C represents the year’s lowest demand point.1 1

To demonstrate this rigorously, let F(y) denote the graph of the load duration curve where y denotes a level of megawatts on the y-axis and F(y) is the corresponding level of hours on the x-axis of the figure. Then the probability that demand is at level y or above is, Prob[demand y] = F(y)/8760 The cumulative distribution function for demand is, G[y] = 1-Prob[demand y] = (8760 – F(y))/8760

8

The California Electricity Crisis: What, Why, and What’s Next

Loads or demand differences are given names based on their proximity to A, B, and C. For example, Peak Demand (Load) are the hours of the year expected to be in the vicinity of A (maximum or peak and demand). Intermediate Demand (Load) are the hours in the year in which demand is expected to be below peak and in the vicinity of B or B' (average or median load). Base-load or Off-Peak are the hours in the year expected to be in the vicinity of the minimum demand, or C. Figure 2-2 shows how a power system’s load duration curve may be segregated using these load level terms.

and the probability density of demand is, g(y) = (y) = - (y)/8760 Median demand occurs at the point, This occurs where, Average demand is given by the area under the load duration curve divided by 8760. To show this we use integration by parts. Average demand is,

Then using integration by parts,

Power Production Economics

9

Power systems satisfy two objectives. First, the costs are minimized. Second, power companies supply electricity to satisfy the varying demands throughout the year as represented by the load duration curve. At any given time, a vertically integrated power company might own various power stations, which might be fueled by different primary energy sources, including: hydro nuclear coal oil and natural gas (steam) oil and natural gas (turbine)

SHORT-RUN MARGINAL COSTS (SRMC) Not all electric utility companies own each type of plant. Fuel availability, fuel prices, environmental constraints, construction and capital costs, etc., determine the plant type (i.e., the mix) owned by specific power companies at any given time. Larger vertically integrated power companies typically own several plants in each category. Within a category, each plant’s marginal cost or marginal productivity might vary considerably based on age, technology, plant design, operating performance, fuel conversion (boiler) efficiency, heat rate, etc. Power systems are operated to minimize the avoidable (i.e., variable) cost of meeting instantaneous load. The necessary condition for meeting this objective is generally expressed using a term known as the “merit order” of generation stations. System operators or dispatchers minimize costs by selecting plants in order of their running or production costs from the least to more expensive.2 As demand is communicated to a power system dispatcher or operator, this simple least-cost operating rule requires that the dispatcher fire up and utilize the plants with the lowest marginal costs. As demand increases, more costly plants must be utilized. These costlier plants generally have: (1) lower heat rates or lower fuel efficiencies; (2) higher fuel costs; (3) higher 2

In practice, marginal generation cost and actual dispatch costs may differ. Actual dispatch cost is based upon a term known as “system lambda” that reflects losses that can occur when energy is transformed by voltage and transmitted over wires. While important in practice, these differences are beyond the scope of this discussion.

10

The California Electricity Crisis: What, Why, and What’s Next

labor costs or other equipment costs; and (4) other locational costs such as line losses or network congestion constraints. As demand increases, more costly plants are called into service (i.e., they are dispatched by the system operator to meet increasing load). Plants with low system “lambdas” are typically called base-load plants and are expected to be called upon nearly 8760 hours of the year. In practice, maintenance schedules and planned, as well as unplanned, outages affect actual use. The annual percentage of time a plant is in use is known as its capacity (or availability) factor (CF). Plants used next in the merit order are only used about half the year. These are known as intermediate plants. These intermediate plants could include older or less efficient plants previously acquired for base-load use that can no longer compete with other base-load plants. Finally, some plants are designed to be used, or their marginal cost makes them economic, for only a relatively small fraction of the year. These are called peaking plants. If marginal operational costs and plant location were the only factors, most power systems would own only low marginal cost base-load plants. This is generally not the case because base-load units generally are more costly to construct than intermediate or peaking plants. This leads to the second aspect of power systems, namely system planning and a related concept, marginal capacity costs.

ELECTRIC POWER SYSTEM PLANNING Power systems evolve to meet growing and changing demand patterns. Power system planners make investments to minimize the present value of future costs (both construction [acquisition] and operational) necessary to meet the forecasted load on their system. Planners generally take as given the plants currently owned and the transmission system already in place. These constraints are relevant, but not absolute. For example, some plants acquired for base-load purposes might be moved down the dispatcher’s merit order and be used for an intermediate or even peaking purpose as new technologies replace older technologies. Investing in new power stations requires considering both a “need” determination and the likely effect a new addition would have on expected fuel and generating costs. As a general proposition, as construction or investment costs increase, operating costs decrease. Nuclear generating stations are relatively inexpensive to operate but expensive to build. Peaking plants are mostly turbines, practically the equivalent of jet engines, and are relatively cheap to acquire per kW of installed capacity. However, a

Power Production Economics

11

peaker’s operating costs are very expensive because they convert expensive fossil fuels to electricity in a relatively inefficient manner. Figure 2-3 depicts how system planners visualize the cost difference of the three primary plant types that might be selected to meet a growth in demand.

The darkened line segments of each cost function form the least-cost envelope for selecting the least-cost plant from each category based upon the number of hours the new plant would be expected to operate. If complex power systems were redesigned and built anew for changing levels or demand, the planning envelope would determine the least-cost mix of plants. In practice, incremental decisions are made that treat the previous investments as givens, plus a reasonable forecast of future needs and costs. Some pragmatic rules emerge. For example, if a new generating station has planned usage between O and B, a peaking plant would be the least-cost addition. If planned usage was above C, a base-load unit would be the leastcost addition. Planning to use a peaking plant beyond B hours would not be cost effective since it would have a higher present value of costs (both construction and operational) than an intermediate plant. And, acquiring an intermediate plant would cease being economic if the station’s use were likely to exceed C hours because, at that utilization level, a new base-load plant would be the most economic choice.

12

The California Electricity Crisis: What, Why, and What’s Next

Additional complications are also important. New plants, once acquired, would likely displace pre-existing plants as the new plants move pre-existing plants down the dispatcher’s merit order. In practice, this means that system planners would consider both the likely fuel savings when new plants could displace less efficient plants currently owned and the costs per kW of the capacity added. To reflect this potential fuel or operational cost saving, a new plant’s construction cost would be replaced by the following: Capacity Cost = Construction Cost – PV(F) where PV(F) = Present Value of Expected Fuel Savings. The least-cost criteria would apply to new investments. Construction costs, carrying costs, and fuel costs would be analyzed jointly to choose the economically efficient generating station. Peaking plants would not have any fuel savings. Over their economic life, base-load plants would typically have very significant fuel savings relative to their construction costs. This effectively moves the intercept of the baseload plants, and perhaps intermediate load plants, down the vertical axis of Figure 2-3. When fuel savings are present, peaking plants, although inexpensive to build, have a higher present value of combined costs. There is a related factor known as marginal capacity cost. Generally speaking, the change in total cost to meet an unexpected increase in demand is equal to marginal energy costs. This holds true because no investment in new capacity is necessary to supply a change in demand when there is sufficient capacity. If capacity must be added to meet increased demand, the change in total cost has two components: marginal energy costs and marginal capacity costs. Time periods where the probability of shortages is high are often called peak periods, while periods with virtually no chance of shortages are referred to as off-peak periods. Vertically integrated electric utilities generally would respond to an increase in demand and insufficient capacity by accelerating the construction or acquisition of a previously planned power station. Ignoring any fuel savings, the annual cost of this acceleration in construction would be the accelerated plant’s amortization cost (i.e., the capital recovery factor (CRF)).3 3

where:

i = the Interest Rate or Cost of Capital and N = Number of Years of expected plant life.

Power Production Economics

13

If there are fuel savings, marginal capacity costs can be expressed as: MCC = CRF[(CC)–PV(F)]

where: MCC

= Marginal Capacity Cost

CC

= Total Construction Costs (Contract Cost)

PV(F) = Present Value of any Fuel Savings due to changes in the merit order of existing plants Typically, a base-load or intermediate load generating station would have lower marginal capacity costs than would a peaking plant, when the fuel savings over the plant’s life are considered. Regardless, additions to power systems in response to an increase in demand affect marginal costs. Pricing is more complex and will be addressed below. In the next chapter we explain how regulatory principles are applied to power systems and discuss the history of cost of service regulation.

This page intentionally left blank

Chapter 3 PRINCIPLES OF TRADITIONAL REGULATION

Electricity supply occurs in three stages. The first stage is producing electricity at the power source. The second stage is transmitting the electricity from the power source to the area where it will be consumed. The third stage is distributing the electricity to the end-use customers. Until very recently, electricity sectors traditionally were vertically integrated, which meant that the three distinct facets of electricity supply were controlled by a single entity. Thus a vertically integrated company typically would own generation, transmission, and distribution assets and resources. Such a company is a monopolist from production to distribution. Such vertically integrated companies tended to focus on the engineering and technological sides of electricity production, transmission, and distribution. Consequently, economies-of-scale1 and rapid technological innovation have, until relatively recently, driven the electricity market. In the previous chapter, we discussed the differences between base-load, intermediate, and peaking plants. Economies-of-scale exist for a vertically integrated electric utility because a large generating system can provide power to many users, and additional users can be accommodated at a comparatively small increase in power costs. In the past, vertically integrated electric companies have been awarded franchise monopoly status in the United States. In exchange for receiving a franchise that allows it to provide electricity on a monopoly basis, a private company agrees to be subject to government regulation. This means that a government regulatory agency will monitor and control prices, profits, and performance. This typically took the form of COS regulation.2 However,

1

Economies-of-scale exist when an increase in inputs in the production process results in a greater proportionate increase in output. 2 Alternatively, the government could own the vertically integrated utility.

16

The California Electricity Crisis: What, Why, and What’s Next

this model has been changing and moving toward a more competitive and efficient structure. Before 1998, California had a mix of privately owned, state-regulated utilities and municipally owned, mostly self-regulated electric utilities. In the 1990s, economists began to question the logic of the industry’s vertical structure and proposed competition to replace the monopoly status of vertically integrated utilities subject to regulation. In California, many critics argued that regulation had failed, as evidenced by rising prices and costly new capital investments. We will discuss these arguments below. But first, we describe COS regulation.

COST-OF-SERVICE (COS) REGULATION U.S. electric utilities are jointly regulated by fifty state regulatory agencies and the Federal Energy Regulatory Commission (FERC). In the past, this joint responsibility has worked reasonably well. The states assumed responsibility for regulating intrastate/retail prices and investments. The FERC assumed responsibility for regulating wholesale and interstate markets.3 Professor Alfred E. Kahn wrote one of the two classic texts on the United States regulatory system.4 He stated: “There are four components of this regulation that in combination distinguish the public utility from other sectors of the economy: control of entry, price-fixing, prescription of quality and conditions of service, and the imposition of an obligation to serve all applicants under reasonable conditions.” 5 These four fundamental regulatory components are based on two basic premises. The first premise is that electricity is an essential good, required for consumers’ well-being. In other words, unlike many other consumer goods that are luxuries, not necessities, electricity is considered to be vital if residences and industries are to function in a manner deemed acceptable for

3

During the California crisis, jurisdictional conflicts between the FERC and the CPUC were prevalent. As the crisis worsened, the CPUC and FERC increasingly blamed each other. 4 Alfred E. Kahn, The Economics of Regulation: Principles and Institutions, Vol. 1 (New York: John Wiley & Sons, Inc., 1970). Volume 2, Institutional Issues, appeared in 1971. The second standard text was written by James C. Bonbright, Principles of Public Utility Rates, 2nd Ed., (Arlington, Virginia: Public Utility Reports, Inc., 1988). 5 Ibid, p. 2.

Principles of Traditional Regulation

17

an advanced society. There is, of course, a fine line between a necessity and a luxury, and the line is not often clear.6 The second basic consideration underlying electric utility regulation is the utility’s monopoly status. As Bonbright noted, electric utilities are generally considered to be “natural monopolies.” “The familiar statement that a public utility is a “natural monopoly” is meant to indicate that this type of business, by virtue of its inherent technical characteristics rather than by virtue of any legal restrictions or financial power, cannot be operated with efficiency and economy unless it enjoys a monopoly of its market. So great are the diseconomies of direct competition that even if it gets an effective start, the competition will probably not long persist, if only because it will lead to bankruptcy of the rivals. Even if the competition is long-lived, as has occasionally happened when the rivalry has taken a restrained form, it is wasteful of resources because it involves unnecessary duplication of tracks, of cables, of substations, and other resources.”7 Regulation was intended to replace competition with administrative restraints on profits. However, this can also be viewed as redistributing income among competing claimants.8 There is a classic tension between the utilities’ claims (and those of its investors) for an aggregate income stream and the ratepayers’ constant clamor for lower rates. To this cacophony, one can add the various customer classes who argue that their rates should be lowered and should not subsidize the rates of other classes. Each class claims to have convincing arguments. Industrial customers argue that a greater share of the rate burden should be passed on to smaller users because the industrial customers’ survival depends on lower rates and that their large and/or interruptible demand imposes lower costs on power suppliers. Residential users seek rate relief for the aged and poor who cannot pay higher rates. Consequently, a landmark regulatory study concluded: “In economists’ language the concern for equity has generally triumphed over the quest for efficiency. In political terms, it means that regulation

6

For example, one can make a strong argument that the first hundred kWHs are necessary for basic lighting and refrigeration needs. The argument for kWHs needed for decorative lighting is less persuasive. Nevertheless, these distinctions are rarely an issue when a utility’s obligation to serve is determined. 7 Bonbright, p. 11. 8 Kahn has described regulation as “the specific charges on different categories of services and the relationship between them.” Kahn, pp. 25-26.

18

The California Electricity Crisis: What, Why, and What’s Next is best understood as a political settlement, undertaken in an effort to keep peace within the polity.”9

In essence, regulatory agencies fear that a monopoly utility, left to its own devices, would charge extortionate or unduly discriminatory rates, or might otherwise exploit customers who do not have an alternative supply source. Consequently, regulators control entry, prices, service quality, and availability. Two conceptual matters are important. The first is that profits or the overall earnings of private companies must be regulated. The second is that the rates charged to individual customers, or to similarly situated customers (grouped into what are known as “customer classes”), must also be regulated. This is called regulating the overall rate level and determining the rate structure, respectively.10 Until recently (and in some places still), U.S. regulatory commissions generally focused on total company revenues.11 Regulators generally first determine operating costs incurred by the utility in a “test” year. To these operating costs, regulators add an allowance for profit. This is usually a specified return on the monies deemed to have been prudently invested in the business. Kahn points out that this involves a three-step process: (1) reviewing operating expenses; (2) determining the investment or “rate base,” and (3) determining the rate of return to be allowed on that rate base.12

Operating Expenses Reviewing a utility’s operating expenses (OE) is generally not a very controversial procedure. Most regulatory commissions have established a uniform system of accounts and try to ensure that a utility’s costs are not excessive or imprudent. Generally, transactions with affiliated companies (e.g., fuel suppliers) are carefully scrutinized. Similarly, salaries are increasingly subject to close review. Utilities are generally required to purchase goods and services at the best prices available. Because competition has been removed, utilities can attempt to pass on costs, even those incurred unnecessarily, to their customers. Thus, regulators perceive a need to subject utilities’ costs to scrutiny in order to reveal inefficiencies.

9

Thomas K. McCraw, Prophets of Regulation (Cambridge, Massachusetts: Harvard University Press, 1984). 10 Kahn, p. 26. 11 Kahn notes that this is “...by far the most hotly contested aspect of regulation...” Kahn, p. 36. 12 Ibid.

Principles of Traditional Regulation

19

This can be accomplished in either an accounting sense (auditing invoices) or in an economic sense (attempting inter-utility cost comparisons).

Valuing Investment It is much more difficult to determine the investment base on which the utility will be permitted to earn a rate of return or profit. This exercise is critical because electric utilities are highly capital-intensive enterprises. Regulators face two controversies in performing this task. The first relates to valuing a utility’s investment, known as Rate Base (RB). While there are many variations on each theme, Rate Base value in the United States can be summarized as the utility’s assets at their original cost, less depreciation.13

Determining Allowable Profits The second and generally most controversial aspect of a regulator’s work comes when he/she must select some profit margin, or “rate of return” (ROR), that will be applied to Rate Base. Usually, the ROR selected falls within a range, bounded by the need to attract capital and by an estimate of what investments in enterprises of comparable risk would yield. The one certainty in all of this is that the ROR will be, at best, an approximation. All utilities employ a mix of debt and equity capital, all have stock prices that are determined by difficult-to-measure anticipated future earnings, and all experience differences between allowed rates of return and those actually earned, the so-called “regulatory lag.” The ROR is typically a weighted average of the interest rate on embedded debt and the authorized rate of return on common equity (ROE).14

Return on Equity Broadly speaking, there are many variants used in the United States for determining the ROE. The most common are: (1) the Capital Asset Pricing Model (CAPM); (2) the Risk Premium (or bare-rent) Method; and (3) the Discounted Cash Flow (DCF) method. The CAPM calculates the systematic risk of a company by comparing either its stock price, or rate of return, with a similar measure for some broad index of stocks (such as the Standard and Poor’s portfolios of stocks). The so-called “beta,” calculated using a regression analysis, is combined with the 13 14

Bonbright, pp. 67, 69, 82-92; Kahn, pp. 35-41. When relevant, the return on preferred debt is also factored into this determination.

The California Electricity Crisis: What, Why, and What’s Next

20

return on a long-term risk-free bond and the risk spread (the difference between equity and risk-free returns, usually the government’s long-term debt) for the portfolio to determine the cost of equity for the utility.

The Risk Premium approach is related conceptually to the CAPM. This method estimates the risk premium between the ROE for a specific utility (proxies are often used) and long-term risk-free debt. The Risk Premium method’s advantage is its straightforwardness. The underlying assumptions and estimates are quite obvious and open to policy review. One could implement this method by estimating the ROE for several industrial sectors. Average industrial risk spreads could be calculated by subtracting long-term government debt from the average industrial return on equity. The resulting average industrial risk-spread estimate could then be adjusted for any perceived differences in risk between the industry in general and the electric sector. The DCF method is based upon what investors expect to earn from the equity that they provide to the utility. In its simplest form, investors purchase stock to earn dividends and to gain from any appreciation in the underlying value of the stock. In the simplified DCF, these objectives are stated as follows:

Revenue Requirements The regulatory COS review then takes the various cost and evaluation factors and determines an annual Revenue Requirement, also called a cost of service. The basic regulatory formula is: RR = COS = OE + RB * ROR

where RR = Revenue Requirement COS = Cost of Service OE

= Annual Operating Expenses, including one year’s non-cash depreciation of past investments

Principles of Traditional Regulation RB

21

= Rate Base of the original cost of all prudent investments

less accumulated depreciation ROR = Rate of Return adjusted for income taxes The ROR needs some additional explanation due to the corporate tax deductibility of interest expenses and corporate taxation of net income. Regulators “gross up” the authorized ROE in determining the ROR used to determine RR or COS. Regulators use the following approach, or something equivalent to determine the tax-adjusted ROR.

where D = Outstanding Debt E = Value of Equity

r = Weighted average interest on debt of various vintages and risk ROE = Authorized Return on Equity t = Corporate Income Tax Rate

Regulated Prices and Tariffs The final step is designing utility tariffs. This step will be addressed in greater detail below. Here, it is sufficient to explain that regulators design utility tariffs that may vary by customer category and uses. In practice, voltage levels matter (e.g., 110 volts for residential and much higher for industrial users). Load pattern differences also matter because different customers have various mixes of peak, intermediate or shoulder, and offpeak consumption. In Chapter 2, we explained how supply costs vary by time of use.

22

The California Electricity Crisis: What, Why, and What’s Next

Although complications in tariff design are very important, the basic idea is straightforward. Regulators must restrict prices so that they produce no more than the utility’s annual revenue requirement. If all prices were equal, and they are not, regulation would simply set the single unit price equal to average revenue. In this case, the regulated price would be:

where AR = Average Revenue Recall that RR is also equal to COS. Therefore, price would also equal:

where ATC = Average Total Cost A conflict arises because, while economic efficiency requires price to equal marginal cost (MC), regulation in effect requires price to equal average cost (ATC). Complexity is added to regulated tariffs, in part, to close the gap between marginal and average cost concepts. In practice, tariff design reflects political, not economic, motivations. This is especially true when costs and prices are increasing. When marginal cost or economic principles are most needed, regulators, for the most part, prefer political redistribution and delay sending bad news to consumers/voters. For many, this inherent regulatory weakness was one of the primary reasons to restructure the electricity industry in California, as it also is in much of the world. In practice, there are separate tariffs for large-volume users (e.g., industrial, commercial, or business consumers) and residential consumers. There are also sub-categories. Historically, politics and income redistribution have affected tariff design and customer cost allocation. During the high-priced 1980s, many larger-volume customers sought relief from what they perceived to be politically motivated cost allocation. These large-volume users added support to the movement that favored deregulating generation, preferring direct retail access through competitive wholesale markets for electricity.

Principles of Traditional Regulation

23

California had regulated retail electricity prices to about twice the prices paid for electricity in the rest of the nation. Industrial users were becoming increasingly tired of paying what they viewed as politically motivated subsidies. Therefore, industrial users joined the forces that pushed to restructure California’s electricity industry by replacing comprehensive costof-service regulation with wholesale competitive generation markets and direct retail access.

Quantity Sold and Forecasting It is incumbent that regulators predict accurately retail sales because volume is the regulatory basis for setting price levels and allocating costs to different customers. Recall that the utility’s revenue requirement is established by adding together the utility’s fixed costs, estimated operating expenses, and related rate base times the required rate of return. The resulting revenue requirement is allocated to customers. Therefore, tariffs must be designed to achieve the authorized revenue requirement. This cost allocation to the various customer groups (and the subgroups within various customer classes) can be, and has been, a very contentious issue in the United States. Adding to the complexity, customers are likely to exhibit price elasticity. However, this demand responsiveness will likely not be the same from one customer class to another. These complex tariff and other policy matters are all tied to the accuracy of the underlying test year’s sales forecast. Ultimately, sales forecasting is an estimate. The estimate’s accuracy depends, in part, on the sophistication applied to the forecast. The level of sophistication can vary greatly across the country. Further, the extent to which economic efficiency is considered a major goal in the development of utility tariffs can have a large affect on how tariffs are set.

Conclusion In the post-Reagan political world, concerns with “pushy” regulation have given political wind to economists’ questions about the natural monopoly status of generation. The growth in successful independent power producers that sold electricity on a MWH, not a power station, basis provided additional impetus to deregulate generation in much of the world. California was at the head of the deregulation pack in several ways. We move on to this matter in the following chapters.

This page intentionally left blank

Chapter 4 RECONCILING MARGINAL COST AND REVENUE REQUIREMENTS

Regulation and competition are mostly different means to the same end. The conflicts between the two approaches are rather obvious. Under regulation, a political body sets prices and authorizes investments, and a monopoly produces and delivers the electricity. Under competition, the market sets prices. It is the interaction between many sellers and generators that establishes a market clearing price (MCP) for electricity. All participants are price-takers in a competitive market. Despite the conflicting premises underlying these two extremes, regulators have attempted to graft marginal cost economic principles onto regulation of prices and investments. This conflict is, at best, a reasoned compromise because revenue requirement constraints are sacrosanct. Nevertheless, the gap is not nearly as great as first impressions may suggest because, as we demonstrated in Chapter 2, marginal cost does play a role for power engineers operating and investing in power systems. In the 1970s and early 1980s, regulators let the power engineers and economists design electricity tariffs based on marginal cost. This resulted in tariffs based on time-of-use price differences. This approach had significant support and eventually fostered the realization that generation was not a natural monopoly. In the 1990s, California and other political jurisdictions determined that competition, not simply better regulated tariffs using marginal costs, was the preferred approach. Today, there are new calls in California to return to the “good old days” represented by time of use pricing that reflects time of day marginal cost differences. This chapter reconciles average and marginal costs. As we discussed above, regulated price levels are based upon collecting revenues to meet a specific annual revenue target (TR), which is typically determined by setting the annual revenue target equal to annual total cost.

The California Electricity Crisis: What, Why, and What’s Next

26

Annual Total Cost (TC) = Annual Operating Cost (VC) plus Annual Fixed Cost (FC) Thus, TC = VC + FC As, annual total revenue is based on annual total cost, we have: TR = TC.

If there were a single product supplied, this would imply that price (P) could be derived by:

where Q = quantity sold, and P is equal to both average revenue and average costs. Under pure competition, a firm in long-run equilibrium receives a price that equals marginal cost (MC). In the long run, a firm would achieve equilibrium by adjusting its investments to earn an adequate or fair return on its investments. Adjusting its scale of operation, the firm would minimize total cost at the point where marginal and overall costs are equal (AC = MC). With perfect planning or foresight, power systems that were allowed to earn a fair return on their investments could achieve the happy long-run result that P = AC = MC.

WHY MARGINAL COST? The logic that would allow regulators to adopt aspects of marginal cost pricing and average cost pricing is predicated on full information. Let’s assume full information and follow the logic. Microeconomics is based on certain precise definitions. On the producer or supply side, a market is considered competitive when there are sufficient competitors so that no firm, or group of firms, possesses market power enabling it to restrict the quantity supplied in the market, thereby influencing the MCP. On the consumer or demand side, a market is presumed competitive when no buyer, or group of buyers, can control the quantity purchased in the market, and thereby, the ultimate price paid. When markets are competitive, the interaction between demand and supply results in an economically efficient quantity sold and price paid. Defining economic efficiency is rather straightforward. When prices are

Reconciling Marginal Cost and Revenue Requirements

27

equal to marginal cost, suppliers get precise signals concerning consumers’ willingness to pay. Suppliers understand that they should produce more when their firm’s marginal cost is less than the market price and should stop expanding output when marginal cost exceeds the market price. Consumers also receive important information. They discover what it costs to supply more units of the good. The price established in the market helps individual consumers determine how much they will consume under the assurance that their individual willingness to pay will equal or exceed the actual market price—and, therefore, the industry-wide marginal cost. In a competitive market, firms supplying product to the market are “price-takers.” These price-taking firms will expand firm output as long as their marginal cost is below the competitive market price determined by demand and supply. And, in a competitive market, firms would reduce their output when their marginal cost exceeds the market price. This is sometimes called operational, or production, efficiency. Competitive markets guide the behavior of the individual firms supplying or producing goods sold. A competitive market’s demand schedule represents conceptually just how much of a good or service all potential consumers would purchase at specific prices (i.e., willingness to pay). In the case of electricity commodity markets, consumers are resellers who purchase power in competitive wholesale markets in order to distribute electricity to retail consumers. If the electricity price is raised, less electricity would be purchased. If the price is lowered, more would be purchased. And, in a scarce market, those consumers willing to pay more would receive the product, not the consumers willing to pay less. This is called allocative efficiency. In wholesale electricity markets, demand is likely to be relatively highly inelastic because purchases are resold and retail consumers may or may not (as was true in California) receive retail price signals that vary in real time. Finally, when competitive markets supply goods that can be stored and consumed in different time periods, interest rates or discounting play an important role in establishing intertemporal efficiency. This concept is also rather straightforward. A dollar today is worth more than a dollar received tomorrow, or next year, or ten years from now. This guiding economic principle affects a firm’s investment, input purchase, and use decisions. This principle also affects when consumers purchase, store, and consume goods. Because electricity cannot be stored, spot markets take on greater importance. Regardless, forward and future contracts, as well as retaining generation ownership, are natural hedges to reduce the risk and cost of spot market volatility. As we will explain later, California’s market design and regulatory practices eschewed such hedging approaches. Instead, most power was sold in what turned out to be highly volatile wholesale electricity markets.

28

The California Electricity Crisis: What, Why, and What’s Next These supply and demand concepts are shown in Figure 4-1.

Price and quantity are the efficient market clearing prices and quantity. The long-run or intertemporal aspects of economic efficiency can be gleaned from reviewing Figure 4-2, where several relevant cost concepts are shown.

Reconciling Marginal Cost and Revenue Requirements

29

Marginal cost (MC) is upward sloping to reflect increasing marginal production costs. Average fixed costs (AFC) decline as output increases over a fixed investment. Average variable costs (AVC) are U-shaped. These costs initially decline and, due to increasing marginal costs, AVCs begin to increase. Average costs (AC) are the sum of average variable and average fixed costs. This curve is also U-shaped, and as AFC approaches zero, AVC and AC approach each other. An important relationship is that MC equals AC when AC is at its minimum point. The same is true for AVC. MC equals AVC when AVC is at its minimum. Some new insight into firm behavior becomes apparent in the long run, where fixed costs, (i.e., investments) can vary. Consider Figure 4-3.

30

The California Electricity Crisis: What, Why, and What’s Next

If the market price falls below the shut-down price the firm cannot cover its variable production costs and should shut down. If the market price equals the minimum AC, the firm is in a break-even position This means the firm will recover both its variable costs and its fixed costs, including a reasonable or fair return on its investment, which is roughly equivalent to interest plus a risk premium for equity. If prices are above the firm would earn a rate of return above “fair” market and attempt to expand (i.e., invest more fixed capital). When prices are above shut-down but below break even the firm is failing to earn a fair return. At this point, it is earning a margin above its AVC and should not stop production. However, the firm will attempt to sell off the assets and shift capital or fixed costs elsewhere, perhaps to other industries. Economists often distinguish between the long run and the short run. The long run is defined as a period in which the firm can alter or vary its fixed cost or investment position. In long-run equilibrium, the firm would not alter its fixed costs (i.e., it would neither expand nor contract its investment position). Therefore, in long-run equilibrium: P = MC = AC

Reconciling Marginal Cost and Revenue Requirements

31

Demand and supply conditions can change in the market and, therefore, perfect foresight is not likely. Thus, a firm would have different long-run equilibria with different future market conditions and price expectations. In Figure 4-4, several short-run average cost curves are shown. Each corresponds to different level of firm investment corresponding to a different expected future market price.

and are short-run average cost (SRAC) curves. and are their corresponding minimums or break-even points. The darker curve formed by connecting the various break-even points is called the longrun average cost (LRAC) curve. The LRAC curve is sometimes called the firm’s planning function because it represents how firm investments would vary in response to different market conditions as an industry grows.

THE LONG-RUN COMPROMISE Long-run equilibrium would be achieved if firms could select their scale based upon the equivalence of price, marginal, and average cost. Under regulation, this is what regulators seek to guide their long-run integrated planning activities.

32

The California Electricity Crisis: What, Why, and What’s Next

Under certainty, the future is known and regulators could logically aspire to this happy synergy, which would allow them to set prices equal to average costs, which would also equal marginal costs. The competitive market would not be necessary to establish prices because, under certainty, the price = AC = MC alignment would be used to set prices. Regulators know they are not prescient. Therefore, they often used complex forecasting models that were subject to much debate and scrutiny. A strong dose of sensitivity analysis was also typically applied. As we explained above, to achieve economic efficiency, a good’s price should equal its marginal cost. All businesses must consider both their short-run costs associated with a given or fixed investment and their longrun investment (or disinvestment) choices. Regulators of public utilities (i.e., the traditional vertically integrated electricity company serving an exclusive franchise) have developed pricing and investment principles or rules that are similar to competitive firm behavior. If forecasts and available information are accurate and these rules are followed, the result is a reasonably good likelihood of economic efficiency.1 The Short-Run Pricing Rule 1. If there is sufficient capacity to meet demand or load, price should equal the short-run marginal running cost (i.e., marginal energy cost adjusted for voltage and losses (time and place)). (Note this is the system lambda described in Chapter 2.) 2. If demand exceeds available capacity, price should equal a higher amount than (1) in order to restrict demand to the available capacity. More precisely, marginal opportunity cost should be added to the marginal running cost.

The Long-Run Pricing and Investment Rule 3. Investments should be made when the marginal capacity cost related to meeting new demand is less than or equal to the short run price that would be charged in (2) to restrict demand to the available capacity without new investments. More precisely, investment should be 1

Several methods exist to allocate common (i.e., fixed) costs across consumers in such a way that the regulated firm breaks even. The fully distributed cost approach, for example, adds a fraction of common costs to the attributable cost of service. While common in regulatory rulemaking, the cost-based approaches are not primarily concerned with economic efficiency or second-best pricing, which is optimal. For further discussion, see Brown and Sibley (1986). Our analysis below emphasizes efficient solutions, known as second-best pricing rules. Stephen Brown and David Sibley, The Theory of Public Utility Pricing (New York, New York: Cambridge University Press, 1986).

Reconciling Marginal Cost and Revenue Requirements

33

expanded if, and only if, the marginal capacity cost is less than the sum of short-run marginal running costs plus marginal opportunity cost. Introducing the concept of marginal opportunity cost, the equivalence of price to restrict demand to capacity and the marginal cost of new capacity can be converted to a marginal cost pricing criteria that applies in all circumstances. Put simply, price should be set equal to marginal cost, including any marginal opportunity costs. If demand is less than available capacity, this means that price should equal marginal production costs (or operating costs, including fuel). If demand exceeds available capacity, marginal opportunity costs need to reflect the likelihood that supply might be insufficient to satisfy demand. This is called a loss of load probability, or LOLP. When LOLP is low, there is virtually no chance that capacity will be insufficient, and vice versa. During peak or high-demand periods, the LOLP would approach unity. Electric prices need to reflect the value of avoiding forced outages or service interruptions. In practice, this often gave rise to an investment rule holding that, if by increasing the price, the market would clear at a lower price than a price equal to marginal production costs plus marginal capacity costs, then prices should be raised to clear the market, but new investments in capacity should not be made. However, investments in new capacity should be made when the sum of marginal running costs plus marginal capacity costs are less than the market clearing price.

COMPETITIVE WHOLESALE POWER MARKET ECONOMICS As we noted earlier, peak periods are the hours when demand presses against the limits of available capacity. Conversely, there is little likelihood of any shortages during off-peak periods. Put in LOLP terms, peak hours have a LOLP close to one and off-peak LOLPs are close to zero. This distinction is the rationale for time of use (TOU) prices, which economists call peak-load pricing. Regulators began to reform their average cost, mostly politically motivated, tariffs in the United States using this type of reconciliation. As we noted earlier, the process that led to this reform uncovered the realization that generation, the primary cost of supplying electricity, was not a natural monopoly. Tariff reform begat market restructuring. This, in turn, led to the California market design and the so-called crisis.

34

The California Electricity Crisis: What, Why, and What’s Next

RECONCILING COMPETITION AND REGULATION2 Economic efficiency generally requires marginal cost pricing. In order to demonstrate this fact, the starting point is maximizing net societal welfare (i.e., the sum of consumer and producer surplus). Let

where P is society’s willingness to pay for varying amounts of Q or output of a particular commodity and f(Q) is the demand function. Welfare (W) is defined as the difference between what society is willing to pay, or total benefits, and the total social costs (SC) of producing the output. The demand function is treated as society’s marginal benefit function. Therefore, total benefits are defined as the integral of the demand function:

Total social costs (ignoring externalities) are also a function of output, Q. The net welfare function can be stated as:

which becomes

where g(Q) is the total social cost function. Maximizing net welfare yields the following necessary condition:

Note that P is the price of the product (such as electricity) and equals the marginal social cost. The above development explains the pricing behavior of a social welfare maximizer and it is, therefore, a normative result. However, pricing rules are usually set in competitive markets. Conversely, regulators may strive to be social welfare maximizers. In competitive markets, the profit-maximizing criteria would replace a welfare-maximizing criteria. The analogous positive formulation is: 2

This discussion can be passed without losing the crux of the analysis in this book.

Reconciling Marginal Cost and Revenue Requirements

35

where is total profit P • Q is the total revenue of a firm, and TC is its total cost, which may or may not be equal to the total social cost. Maximizing profits yield the following necessary condition:

If

(price elasticity of demand)

and

(marginal private costs)

then, this necessary condition can be restated as follows:

and dividing both sides by P yields

This formulation is sometimes referred to as the Lerner degree of monopoly power.3 It measures the extent to which noncompetitive firms will mark up price relative to competitive firms. In a perfectly competitive world, each firm would be a price-taker and would be negative infinity. Then is zero, and the profit maximizing pricing rule would be identical to the normative case, namely that P = MC. Assuming that marginal private costs (MC) and marginal social costs (g'(Q)) are equal, competition and regulation would both seek to set prices equal to marginal cost. Abstracting from externalities, as price elasticity increases from negative infinity (or being perfectly elastic) to become less elastic, the firms that can set or influence prices would mark up their prices relative to marginal costs. 3

For a discussion of Lerner’s markup equation, see S. Weintraub, Intermediate Price Theory, (New York: Chilton Books, 1964).

36

The California Electricity Crisis: What, Why, and What’s Next

REVENUE CONSTRAINTS The implication of a utility’s revenue constraint is important. Our analysis follows the pioneering work of Baumol and Bradford.4 However, the key results were actually obtained much earlier by Ramsey (1926).5 Assume that the revenue constraint is set exogenously and does not depend upon output, (Q). Also assume that a regulatory commission sets net allowed income equal to (R). Assume further that (a) these are n different customer types (large industry, business, residential); (b) each class has a different price elasticity of demand; and (c) that the utility has different costs for supplying these different customer types. (Homogeneous demands and costs are assumed within each customer category.) Welfare will be defined under these conditions in a manner analogous to (3). Assuming, for simplicity, that demands are independent and costs are fully separable across customer types, then total benefits equal the sum of demand functions across customer types and total social costs equal the sum of the total social costs of supplying each customer type. Therefore, the refined welfare function can be defined as:6

where is the demand or willingness to pay function for the ith customer type, and

is the social cost function of supplying the ith customer type. A regulatory constraint (R) that fixes allowed regulatory profits can be added to the analysis. Let:

where

4 5

6

W.J. Baumol and D.F. Bradford, “Optimal Departures from Marginal Cost Pricing,” American Economic Review, Vol. 60, June 1970. F.P. Ramsey, “A Contribution to the Theory of Taxation,” Economic Journal, 37, (1927), pp. 47-61. Derivations for the case of non-independent demand are straightforward. See Brown and Sibley (1986) and Kenneth Train, Optimal Regulation–The Economic Theory of Natural Monopoly, (Cambridge, Massachusetts: MIT Press, 1995).

Reconciling Marginal Cost and Revenue Requirements

37

TR = total revenue = TC = total cost, which for simplicity can initially be assumed to be set equal to total social costs (SC) (so that

and

R = allowed regulatory profits The constrained objective function then becomes:

Maximizing W* yields the following set of necessary conditions:

for all i = 1, n

where is the Lagrange multiplier and can be interpreted as the marginal welfare loss (gain) due to the regulatory constraint. It is also known as the Ramsey number. If

is the price elasticity of demand for the ith group, then

the necessary condition for each ith customer type can be reformulated as follows:

The California Electricity Crisis: What, Why, and What’s Next

38

Dividing by

yields:

Rearranging terms further and substituting in for yields:

for elasticity and

or the marginal social cost of supplying the ith customer group

The regulatory constraint on income causes the utility to deviate from unconstrained marginal cost pricing.7 In the normative formulation, such as this, the R constraint causes the loss in welfare to be spread over each customer group in a manner that results in the least loss in total welfare for all customers collectively. Ramsey pricing is a second-best solution. To understand the mark-up rule a bit better, assume that a firm with fixed costs and increasing average cost sets price equal to marginal cost. In this case, the firm will fail to break even. The Ramsey pricing rule modifies price a little over marginal cost in such a way as to perturb the first-best outcome as little as possible. To do this, markets with the least sensitivity to demand experience larger price increases, but the quantities purchased are not affected to a large degree. This strategy allows fixed costs to be covered while staying close to the first-best price equals marginal cost solution. Two cases are important. A utility may earn too much revenue because costs are increasing and charging all customers their marginal cost may produce a surplus for the firm that exceeds R. Alternatively, costs may be decreasing; therefore charging all customers their marginal cost will fail to recover an income of R. Turvey points out that revenue constraints need not depend on the presence of increasing or decreasing costs. He explains that: “Marginal cost pricing, subject to any appropriate adjustments for nonoptimalities, may yield a revenue which provides a surplus in relation to accounting cost which is too high... or too low.... If this financial constraint is sacrosanct, then some prices will have to be set below 7

Note that if the revenue constraint is not binding, then the associated Lagrange multiplier is zero and will equal

Reconciling Marginal Cost and Revenue Requirements

39

marginal costs (in the first case) or above (in the second). In either case a constraint which is effective will result in a welfare loss.”8 The different policy implications of increasing and decreasing costs are important because increasing marginal costs are likely in competitive markets, while natural monopolies are essentially decreasing cost industries. Equation (15') helps to understand these two cases. Under marginal cost pricing, and would be the price charged to and quantity consumed by the ith customer type. If the resulting income (R) is excessive, then the regulatory constraint requires a lower price and greater quantity. This situation occurs under marginal cost pricing. The shaded area in Figure 4-5 represents the welfare loss from regulation for the increasing cost case. In order to understand the limits on it is useful to restate the Lerner monopoly, perfectly competitive, and regulatory pricing rules; when and are identical. Monopolist: Competitive: Regulatory:

8

R. Turvey, Optimal Pricing and Investment in Electricity Supply (Cambridge, Massachusetts: The MIT Press, 1968), p. 89.

40

The California Electricity Crisis: What, Why, and What’s Next

When marginal cost increases, as in Figure 4-5, monopolists would mark up price above marginal cost, while regulation would require prices to be set below marginal cost. Therefore,

where = regulatory price = marginal cost price = discriminating monopolist price From (16) and these pricing rules, it follows that when marginal costs are increasing:

This also means that when costs are increasing: (1) no customer should be charged a regulated price that exceeds the marginal cost of supplying that customer, and (2) the percentage price decrease should be greatest for those

Reconciling Marginal Cost and Revenue Requirements

41

customer groups that have the most inelastic demand.9 Finally, regulation in this case imposes a net welfare loss relative to marginal cost pricing. The second case is the decreasing or classic natural monopoly cost case of regulation. Different limits of and orderings of price are applicable under the decreasing cost case. Consider Figure 4-6. When costs are decreasing, pricing based upon marginal cost will produce too little income. Under regulation, prices will need to be set above marginal cost. The regulatory constraint has a similar effect, increasing prices, as a monopoly. Regulation, however, prevents the excesses of unrestricted monopoly pricing. Therefore, regulated prices will be below those that a monopolist would charge. The shaded area in Figure 4-6 represents the welfare lost due to regulation.

The above pricing rules establish the following orderings of prices and for the decreasing cost case:

and, therefore:

and 9

must be positive:

This is the classic Ramsey inverse elasticity rule.

42

The California Electricity Crisis: What, Why, and What’s Next

Note (17") means that when costs are decreasing: (1) no customer should be charged a price that is less than the marginal cost of supplying him, and (2) following Baumol and Bradford, the percentage price increase should be greatest for the most price inelastic group. However, the regulatory price markup for any customer group should be less under regulation than the price markup that a monopolist would add. The loss in welfare and quantity would also be less.

PEAK LOAD PRICING The most important extension of the previous analysis for utilities for peak and off-peak demand differences is Oliver Williamson. 10 A welfare function similar to that utilized above can be used to determine optimal pricing rules when both capacity costs and operating costs vary for peak and off-peak times. It is a straightforward task to extend the previously determined pricing rules for this case.11 Williamson defines capacity costs using the amount of time over a year that any customer utilizes capacity and the percent of time that capacity is fully utilized. It is simpler to consider on-peak and off-peak as two separate services. Thus, off-peak costs are defined only in terms of operating costs and on-peak costs in terms of both operating and annual capacity costs. Since demands are assumed to be independent, a separate demand schedule can be established for each customer group for on-peak and off-peak use. Using this simplification for electricity, off-peak prices would be set equal to marginal operating or running costs off-peak. On-peak prices would be set equal to the marginal operating or running costs on-peak, plus the marginal capacity costs.

10 11

O. Williamson, “Peak Load Pricing and Optimal Capacity Under Indivisibility Constraints,” American Economic Review, LXI, Sept. 1966, pp. 810-827. Several complications such as shifting peaks, marginal opportunity costs, and interdependent demand and supply are ignored here.

Reconciling Marginal Cost and Revenue Requirements

43

For simplicity, let there be two time periods, which are peak (p) and offpeak (o). Further, let there be only two customer groups, 1 and 2, and assume that customer 1 consumes only off-peak and customer 2 consumes only on-peak. The objective function is:

where are the respective areas under the demand schedules for 1 and 2 is the off peak variable cost and a function of is the on-peak cost and has two components: operating cost during peak periods

and

is the variable or is the annual

capacity cost that is incurred to maintain sufficient capacity to supply peak period demands thus related directly to short run opportunity costs. Maximizing with no interdependence or shifting peaks yields the following necessary conditions:

where is the marginal operating costs of supplying off-peak is the marginal operating costs on-peak is the marginal capacity costs amortized for on-peak use

Equations (20) and (20') may be readily generalized. It is a simple matter to generalize these results by altering demand to high-demand on-peak uses (typically weather-related) and off-peak uses, such as base usages, when economic activity and weather conditions cause low demand. Put this way,

44

The California Electricity Crisis: What, Why, and What’s Next

off-peak means little chance of a significant LOLP. On-peak means that LOLP would approach one. The short-run pricing rule (price equal to marginal cost) is accommodated conceptually since marginal capacity cost is assumed to be less than or equal to the short-run marginal opportunity cost of avoiding shortages or outages. is the price elasticity of demand on-peak = is the marginal regulatory effect These results can be further extended to include interdependencies in demand and supply, 12 as well as any regulatory incentives that would encourage utilities to over-invest when regulation bases authorized net income (R) on rate base of the amount invested.13

NON-UNIFORM PRICES Ramsey prices are a set of uniform prices that maximize total surplus (producer and consumer) subject to a break-even or allowed revenue constraint. Ramsey prices do this by setting different prices in various markets (customer classes or time-of-day). Another solution to the breakeven constraint in public utility regulation is the use of non-uniform prices. Non-uniform prices allow different customers within a market to pay different prices. Typically, this is done by tariffs that have multiple tiers or entry fees. Optimal tariff design is beyond the scope of our analysis, but is fully reviewed in Brown and Sibley (1986).14

CONCLUSION The previous discussion adds mathematical complexity to reconcile and accommodate economic principles, regulation, and optimal investment. As 12

See R. Turvey, Economic Analysis and Public Enterprises, (Totowa: New Jersey, Rowman and Littlefield, 1973); R. Rees, “Second-Best Pricing Rules for Public Enterprise Pricing,” Economica, August 1968. 13 See E. Bailey, “Peak Load Pricing Under Regulatory Constraint,” Journal of Political Economy, Vol. 80, July/August, 1972; H. Averch and L.L. Johnson, “Behavior of the Firm Under Regulatory Constraints,” American Economic Review, Vol. 52, 1962; S.H. Wellisz, “Regulation of Natural Gas Pipeline Companies: An Economic Analysis,” Journal of Political Economy, February 1963, pp. 30-43. 14 Brown and Sibley (1986), pp. 61-97.

Reconciling Marginal Cost and Revenue Requirements

45

the pricing rules become increasingly complex, it is not surprising that there is little reasonable likelihood that regulators would have the necessary data and precise inputs to establish utility tariffs with precision. In the 1980s, when regulators were attempting to implement many of these conceptual ideas, a common refrain was, “It’s better to be roughly right than precisely wrong.” Despite such brave assertions, regulation could only strive to achieve efficient pricing using simplified tariffs. Competitive wholesale markets were viewed as a better means to the end because realtime prices could signal the short-run value of avoiding shortages. Further, in a competitive wholesale market, the economic rents earned (prices above marginal running costs) would signal new investments in generation and cause new competitive entry if the economic rents were sufficiently large relative to competitive rates of return. In the 1990s, efforts to restructure regulation and utilities and to move toward competitive wholesale markets began to take root. Real-time or time-of-use (peak/off-peak and more) pricing was set using competitive bidding schemes between sellers and buyers. The concept was correct. It was the execution that failed in California.

This page intentionally left blank

Chapter 5 COMPETITIVE WHOLESALE MARKETS FOR ELECTRICITY

This chapter outlines the general concepts used to design competitive wholesale electricity markets. The book’s partial purpose is to explain what happened in California. In the following discussion, the rudiments of the California approach will be used to introduce basic concepts. It is important to note that no single restructuring approach has been used around the world to change vertically integrated monopolies into competitive wholesale generation markets. The FERC’s difficulty in getting political acceptance of a Standard Market Design (SMD) unintentionally proves the point that few political jurisdictions believe that “one size fits all.” Doubtless, experimentation and shared lessons learned will prevail. In this context, “what went wrong” in California takes on more than a passing interest.

MARKET DESIGN PRINCIPLES California was an early entrant into electricity market restructuring. For the first two years, it had a rather successful restructuring experience as wholesale generation prices averaged about half of their previous cost of service levels. In order to restructure electricity industries, one generally needs both a market clearing entity and a network reliability provider. In California, the basic competitive structure encompassed the California Power Exchange (CPX) and the California Independent System Operator (CAISO). The CPX met the former requirement and the CAISO met the latter. Figure 5-1 shows the physical flow of power under the California model. Assume that there are N generators. Each bids to sell electricity (kWs) at various hours in the year at different prices. The CPX selects the lowest bid

48

The California Electricity Crisis: What, Why, and What’s Next

price to satisfy the increment of power needed by consumers each hour in the year. The winning generation bidders were then paid the CPX market clearing bid price through the CPX billing/settlement protocols, which charged consumers this same price. Thus, the CPX established market clearing prices (MCPs) every hour using one day ahead bidding. The CAISO operates the electricity network and has responsibility for network reliability. If congestion or network reliability is threatened, the CAISO alters the dispatch schedule using real time market bids and zonal factors. The CPX and CAISO markets were sequential, but not integrated operationally. Ongoing monitoring was parochial in nature. The CPX monitored its market, but not the CAISO’s market. The CAISO monitored its own market, but not the CPX’s market.

Some customers (e.g., would purchase directly from the CPX and receive delivery over the transmission grid operated by the CAISO. Other customers (e.g., would use a broker or aggregator to purchase kWs. These customers would pay the broker, who in turn pays the CPX. The CAISO would assure the delivery of the kWs to the customer. And, some customers (e.g., would purchase electricity directly from a distribution company (most likely the predecessor utility, The distributors would buy either directly from the CPX or from a broker/aggregator (e.g.,

Competitive Wholesale Markets for Electricity

49

Tables 5-1, 5-2, and 5-3 outline the nature of the decision-making in a reasonably competitive wholesale market for sellers (generators), buyers (customers, brokers, etc.), and the power exchange.

Figure 5-2 shows the merit order of the selling bids made by generators at a particular hour and date. The CPX would rank these bids in order from lowest to highest. Figure 5-3 shows the amount that buyers would bid to pay for different quantities of power at the same date and time. Finally, Figure 5-4 shows how the CPX would compare the bids of the sellers and bids of the buyers to establish the MCP and which bids to process.

50

The California Electricity Crisis: What, Why, and What’s Next

Competitive Wholesale Markets for Electricity

51

Table 5-4 summarizes the transactions that would occur in this particular competitive power market.

Conclusion Competitive wholesale market approaches for organizing electricity markets are based upon the conceptual principles of competitive markets. For markets to meet competitive criteria, it is necessary that neither generators nor buyers have market power and, thus, cannot establish price or quantity

52

The California Electricity Crisis: What, Why, and What’s Next

outcomes. California encouraged new generation ownership to reduce market concentration on the supply side. Incumbents were encouraged to divest existing generation and to stop building new generation. On the demand side, California encouraged new energy service providers (ESPs) to compete against the incumbent utilities. Initially, supply exceeded demand and most observers failed to perceive that these initial attempts to reduce market power: (a) were not sufficient and/or (b) failed to account for other design flaws. Further, no one recognized that the elements of the “Perfect Storm” were looming on the horizon.

Chapter 6 CALIFORNIA’S MARKET DESIGN: AN INITIAL SUCCESS FOLLOWED BY A “PERFECT STORM”

California has had expensive electricity for several decades. During the 1980s and 1990s, California’s retail electricity prices were about twice the national average. The general political perception in the mid-1990s was that California’s businesses could not compete given the state’s extraordinarily high electricity prices. Governor Wilson and the state’s legislature looked for things the state could do to help revive and sustain the California economy. Electricity prices became a focal point. The conventional wisdom at the time was that CPUC regulation was too “pushy.” The CPUC was also blamed for allowing the utilities to build expensive nuclear plants and purchase power from independent power producers that sold environmentally friendly electricity. In addition, the CPUC foisted conservation and other social costs on the utilities. All these higher costs were passed on to retail customers, and many also found the regulated tariffs were unfair. Californians were convinced that traditional COS regulation had failed and that a new approach was needed. Ultimately, California restructured its electricity industry using a competitive wholesale generation market approach based on the concepts discussed in Chapter 5.

HOW DID CALIFORNIA REACH THIS POINT? California’s high electricity prices were caused by a combination of bad luck and some CPUC failures. First, California had no in state coal-fired generation plants: The state’s air pollution restrictions and lack of coal

54

The California Electricity Crisis: What, Why, and What’s Next

deposits made it impossible to generate electricity using coal-fired generation in California. Most of the nation had lower electricity prices than California, in part, because these other regions were able to use relatively low-cost coal to produce electricity. Second, federal legislation enacted subsequent to the two oil crises and natural gas shortages in the 1970s made it impossible for California to build new oil or natural gas-fired generation. This meant that during much of the 1970s and 1980s, California had no in-state fossil fuel generation choices. Third, other than some limited hydroelectric generation, California had but one new generation option: nuclear power. However, this generation option was capital intensive. Severe inflation and high interest rates subsequent to the two oil crises in the 1970s made the nuclear power option almost too expensive for regulators and consumers to tolerate. Nevertheless, with no other viable supply-side options, California did invest in expensive nuclear generation plants. These enormous nuclear investments led to higher retail prices. California subsequently made the regulatory/political choice not to fall victim to such costly electricity generation again. Regrettably, California regulators chose to subsidize conservation and energy efficiency, while also agreeing to require the state’s IOUs to pay what seemed at the time to be exorbitant prices for renewable energy under the “must purchase” terms of QFs, an option that itself was extremely expensive. With no new generation being built in the state, California also became dependent on out-of-state coal, nuclear, and hydroelectric generation from the rest of the western region, incurring expensive transmission fees in the process. Nevertheless, in normal weather years, this arrangement, particularly hydroelectric imports, benefited California because the state could import hydropower in warm months (when it needed power) and could export surplus power to the Northwest in cooler months (when California did not need power). Thus, Californians paid twice as much for power due to: (1) environmental restrictions; (2) expensive transmission investments; (3) expensive nuclear power; (4) subsidized conservation, efficiency, and renewables programs; and (5) costly QF power-purchase contracts. In sum, this represented a profound failure for traditional COS regulation. These circumstances led to restructuring the IOU portion of California’s electricity industry. A collaborative process was used to restructure whereby many parties came together to redesign the IOU industry in California. The municipally owned utilities (MOUs) were invited to join the process, but the MOUs opted out. After designing the new system, the parties took the complex package to the state legislature, where a new law entitled AB 1890 was enacted. AB 1890 has now become infamous in California and among energy market participants around the nation and world.

California’s Market Design: an Initial Success Followed by a “Perfect Storm”

55

AB 1890 contained four principal components. A four-year transition period was established to give the IOUs a reasonable opportunity to recover any uneconomic, and therefore potentially stranded, costs. The utilities were also permitted to impose a competition transition charge (CTC) to recover stranded costs. IOUs were urged to divest most of their in-state fossil fuelfired generation. Finally, the IOUs were required to cede operational control over their transmission assets to the CAISO. Two new institutions were established: the CPX and the CAISO. As we explained earlier, the CPX established wholesale spot power markets while the CAISO was responsible for system reliability and transmission congestion management. A single MCP was adopted. The CAISO was designed to alter dispatch when and if system congestion or reliability problems arose. Virtually all the electricity used to supply the IOUs’ customers was required to be sold through spot markets at a single MCP. During the first two years, as projected by those who designed California’s restructuring, supply exceeded demand and the weather cooperated. Wholesale power prices were about one-half (roughly $25/MWH) of their previous cost-of-service costs (about $50/MWH).1 The state’s energy experts expected that at the end of the four-year transition period, California’s retail prices would decrease by an additional twenty-five percent. Energy analysts and regulators flocked to California to learn how they too could import the Golden State’s restructuring miracle to their own jurisdictions. Investors were also quite satisfied as they learned that the state’s three IOUs would likely recover all their potentially stranded costs before the retail price freeze was scheduled to end in April 2002. The only unhappy people in the mix seemed to be the new owners of California’s incumbent generation who sustained some regulatory and wholesale market setbacks when wholesale prices went soft and onerous after-the-fact regulated must-run tariffs were imposed on certain of their newly acquired generation assets. Restructuring in California failed in its third year when a “perfect storm” shocked California. The storm hit just as plans were being finalized to end the four-year transition cost recovery earlier than expected because the three IOUs had recovered their stranded costs more quickly than had been projected. Three things happened in the third year, each of which, on its own, would have caused competitive wholesale power prices to surge. When all three occurred simultaneously, California’s restructuring efforts and the new markets collapsed.

1

Generation represents about half the retail price.

56

The California Electricity Crisis: What, Why, and What’s Next

A PERFECT STORM Three factors combined to form this “perfect storm” in the west. First, economic expansion in California outpaced forecasts. Consequently, in 2000, demand overtook and overwhelmed supply. Most analysts had expected demand to catch up to available supply in 2002 or 2003. Accordingly, new generation licensing and construction was based upon in service dates well beyond 2000. Second, a world oil price surge was coupled with a rapid run up in natural gas prices in the fall of 2000. In fact, by December 2000, North American natural gas prices were more than 500 percent greater than they had been a year earlier. California’s electricity prices are predominantly tied to natural gas because California’s marginal generation units, meaning its most expensive and last utilized power stations, are mostly natural gas-fired. Third, 2000 and 2001 were marked by anomalous weather patterns in California and the Pacific Northwest. Contrary to normal weather patterns in the west, the Northwest was dry when the south was hot. Typically, the western part of North America is either “dry and cool” or “wet and hot.” The unusual weather pattern experienced in the 2000/2001 time period was the worst possible combination for a generation-short region like California because, just as summer demand surged, available hydroelectric supply fell off as drought in the Pacific Northwest dried up hydro power sources. The swing relative to a normal year was equivalent to nearly a twenty percent loss of California’s IOU-owned generation requirements (about 8,000 MWs out of 40,000 MWs).

DEMAND/SUPPLY IMBALANCE We need to explain in greater detail the combined effect of these three forces to fully convey what happened in western energy markets in 2000 and 2001. When the California Legislature deregulated the state’s wholesale electricity markets in 1996, the state and the whole western region had approximately 20 percent excess generating capacity. Adhering to economic theory, this excess supply kept California’s competitive market’s wholesale electricity prices relatively low. However, the experts were surprised by California’s rapid recovery from the recession. Fueled by a hot regional economy, regional demand for electricity grew rapidly in the late 1990s. This economic success is evidenced by growing personal income. In 2000, California’s personal income growth rate (9.3 percent) outstripped the personal income growth rate in the western region (8.9 percent). Personal income growth means an increased demand for goods and services. As electricity is a major factor in

California’s Market Design: an Initial Success Followed by a “Perfect Storm”

57

producing those goods and services, increased demand for goods and services leads to an increased regional demand for electricity. The western region’s population growth also outstripped the national average. From 1996 through 2000, the western region’s population grew at 2 percent per year, compared to the national 1.5 percent average. California grew at an even faster 4 percent per year. Increases in population mean rising electricity demand because more people and more households require more electrical service. However, while demand was growing, there were no significant centralized generation stations built in California between 1996 and 2000 to meet this growing demand. In fact, only three small generating plants were built in the two years prior to the markets opening in 1998. Figure 6-1 shows that the additions to electricity supply were inadequate to meet increased peak demand in California and the western region. This failure to keep pace with rising demand was a key factor in California’s dramatic wholesale electricity price explosion beginning in the late spring of 2000.

Generators did not site and build sufficient new power generation to keep pace with demand in the years before 2000 in either California or the region. As a result, the excess energy available when deregulation took effect in 1998 had all but disappeared by the summer of 2000. The increased demand

58

The California Electricity Crisis: What, Why, and What’s Next

and decreased supply contributed to California’s unprecedented wholesale electricity price surge.

WEATHER The wholesale electricity markets in the western region of the United States are interdependent. Power typically flows between the states in the winter and summer months when certain states are cold and others are warm. During the summer months, the Northwest typically sends power to the Southwest. In the winter months, the reverse is generally true. By coordinating the supply of and demand for electricity, each state can optimize its generation, thereby reducing the overall need to build power plants in the western region. Unfortunately, weather patterns were not normal in 2000. There was an extended drought in the Northwest and the Southwest experienced extremely hot temperatures. Figure 6-2 shows that the water runoff for the Northwest in the first seven months of 2000 was below the 30-year average. It was also less than any year since 1995. Rain and snow fuel the Northwest’s hydroelectric generators, which produce electricity to export to the Southwest. The drought meant that 8,000 to 12,000 MW of electricity were not available to be sent to the Southwest in the spring and summer of 2000. This meant that 15 percent to 20 percent of the electricity California needed for the late spring and summer was suddenly and unexpectedly unavailable from traditional out-of-state sources.

California’s Market Design: an Initial Success Followed by a “Perfect Storm”

59

As we noted above, the drought in the northwest coincided with earlierthan-normal hot regional weather in the spring.2 This hot weather increased air conditioning demand. Figure 6-3 shows the average CAISO demand for electricity in California in May and June 2000. As shown in Figure 6-3, demand in May and June 2000 was significantly greater than during the previous two years, due in part to the unusually hot weather.

The unusual weather patterns continued unabated throughout the summer of 2000. California recorded average temperatures for May and June 2000 that were the hottest on record in over 80 years, July was the hottest in over 20 years, and August was the hottest in more than 60 years. This pattern held true for the entire Southwest, where May through August 2000 were ranked among the hottest in more than 80 years. The Southwest’s hot weather (causing increased demand) and the Northwest’s low water runoff (causing decreased supply) were major factors that affected the wholesale market for electricity in California.

NATURAL GAS PRICES There are approximately 1,000 power plants in California that can produce approximately 53,000 MW of electric power. Approximately 325 of the state’s power plants use natural gas. Natural gas-fired capacity constitutes 2

May and June 2000 ranked among the 15 hottest May/June periods in the western region in the last 100 years.

60

The California Electricity Crisis: What, Why, and What’s Next

52.3 percent, or 27,829 MW, of the state’s total generating capacity. Natural gas plants are often preferred in environmentally sensitive California because they typically emit fewer pollutants than plants that run on coal or diesel fuel. Prior to 2000, California’s natural gas prices had been low, which made California’s gas-fired plants relatively inexpensive to operate. As weather conditions reduced the hydroelectric supply to historically low levels, California was required to run in-state generation more often, thus burning more natural gas. Between May and September of 2000, natural gas consumption by electric utilities in California was 22.4 percent greater than it had been for the corresponding months in 1999. In the western United States, electricity generation demand for natural gas increased by 62 percent during this period. These western regional weather-related problems and natural gas consumption requirements were exacerbated by an explosion on the El Paso Pipeline system, which interrupted natural gas deliveries to Southern California in the summer of 2000. This interruption continued throughout the entire crisis in California. As late as October 2001, capacity on the system into Southern California was still down 20 percent. The pipeline explosion and supply interruption dramatically decreased summer storage operations and curtailed fall deliveries. As a result, California IOUs, especially SCE, entered the winter of 2000 with natural gas storage levels that were far below normal. Natural gas can generally be purchased under delivery contract or on the spot market. Without a delivery contract, electricity generators buy their natural gas supplies on the spot market, where the price is subject to greater fluctuation and includes a convenience and availability surcharge. In a single price competitive electricity market such as the one in California, spot prices for natural gas become important because the generating station that typically sets the MCP is gas-fired. Increased spot natural gas prices increase the marginal running costs, and thus increase the MCP. Through June 2000, natural gas spot market prices were about $3 per million cubic feet (MCF), or less. As shown in Figure 6-4, during the fall and winter of 2000, spot market prices surged. In December 2000, spot natural gas prices jumped more than twenty fold to about $60 per MCF and at times exceeded $60 per MCF in mid-December 2000. According to the Energy Information Administration, the increase in natural gas prices was due to the inability of production to keep pace with the rapid growth in demand, which was driven by the shift toward natural gas as a means to fuel electricity production.

California’s Market Design: an Initial Success Followed by a “Perfect Storm”

61

The sudden and sharp rise in natural gas prices had an immediate effect on California’s wholesale electricity prices. Consider the fact that a fairly efficient, natural gas-fired power plant consumes approximately 10 million British thermal units (10 MMBtu) of natural gas to produce one megawatt hour (MWH) of electricity, and further assume that the wholesale spot price for one MWH of electricity was $20. If natural gas prices rise from $3 per MMBtu to $4 per MMBtu, the increased fuel cost alone would raise the marginal running costs from $30 per MWH to $40 per MWH. However, the actual increases in spot natural gas prices in California in 2000 were much greater. At $30 per MMBtu, marginal running costs would increase to $300 per MWH. If sellers bid their marginal running costs, $300 per MWH would be the MCP. Under California’s single price auction, all sellers would be paid $300 per MWH, including those generators with marginal running costs lower than the $300 per MWH MCP. While this outcome is efficient, it caused a significant increase in IOU purchase power costs.

AIR QUALITY RESTRICTIONS California is environmentally sensitive. The state’s Air Resources Board develops air pollution standards to implement federal and state air quality regulations and standards. Power plants must comply with these strict standards or pay expensive fines. Each plant is issued a certain number of credits annually that entitle it to emit a finite amount of pollutants, including nitric oxide and nitrogen oxide (NOx). One credit is equal to one pound of emissions. A good rule of thumb is that a typical natural gas-fired plant will emit one or two pounds of NOx for each MWH produced.

62

The California Electricity Crisis: What, Why, and What’s Next

Exacerbating the crisis in California, NOx reclaim credits become much more expensive. Once a plant uses up its allocated credits, it must either be shut down or additional credits must be purchased. In January 2000, a NOx reclaim credit sold for about $1 per pound. By May 2000, the price had risen to $5 per pound. By December, NOx reclaim credits were fetching up to $46 per pound.3 Assume that a power plant in California emits two pounds of NOx per MWH produced (a typical level). Had the plant operator purchased NOx reclaim credits in August 2000 at about $30 per credit, the cost of these credits alone would have added $60 (2 x $30) to the cost of producing each MWH. Thus, a MWH that cost $35 to produce in January 2000 would have cost $95 to produce in August, based solely on the increased price of NOx credits. By December 2000, NOx credits were adding nearly $100 per MWH to the cost of generating electricity for the marginal units. When one recalls that $100 per MWH is equal to 10¢ per kWH, which is about twice the delivered price of retail electricity in much of the United States, the gravity of California’s restructuring crisis comes into sharp focus.

CONCLUSION As California’s aging natural gas-fired plants were called upon to run much harder and longer than anyone had anticipated, NOx reclaim credits became scarce and more expensive. Operators often chose to shut down power production if credits were not available or were extremely high priced.4 Whether plants were curtailed or shut down, the outcome was the same: reduced supply. Reduced generation supply, in turn, caused wholesale market electricity prices to increase. With diminished supply, if demand does not change, the price will increase. Therefore, whether plants chose to purchase high-cost reclaim credits (adding to marginal running costs) or curtail production (decreasing supply), the surge in NOx prices also affected wholesale electricity prices in California, contributing to the sharp price escalation. Thus, these uncontrollable market factors collided, creating the “perfect storm.” Without a doubt, these market factors contributed to the rapidly escalating prices that roiled California’s electricity market. Next, we examine how design flaws exacerbated and contributed to the crisis. 3 4

FERC Staff Report dated November 1, 2000; CAISO. During 2000, there are indications that several natural gas-fired power plants had to curtail their output because they had exhausted their available NOx reclaim credits.

Chapter 7 DESIGN FLAWS AND A WORSENING CRISIS

During the first two successful years of restructuring in California, prices declined. This initial success meant that the restructured market’s design flaws were mostly overlooked by market participants. However, in 2000, these design flaws became very significant. The collaborative restructuring process built a market design with some questionable design components. These include: The CPX and CAISO created multiple sequential markets that encouraged strategic bidding. IOUs were required to buy and sell virtually all electricity in the spot market. The CPUC limited long-term purchase power or hedging contracts. The CAISO was required to purchase electricity outside of market (OOM) at prices above California’s market caps to ensure reliability. The CPUC retained the freeze on retail electricity prices even as wholesale prices surged. The CAISO was unable to control or coordinate scheduled electric generation outage schedules. Below we will discuss each of these flaws and explain how each contributed to the California electricity crisis.

Sequential Markets and Strategic Bidding Under AB 1890, a new entity know as a Scheduling Coordinator (SC) was established. The SCs were responsible for arranging generation dispatch, transmission, energy, and capacity for market participants. SCs were

64

The California Electricity Crisis: What, Why, and What’s Next

required to be certified by the CAISO, paid CAISO’s charges, submitted weekly and annual forecasts, day-ahead and hour-ahead schedules, and paid for non-self provided ancillary services. The CAISO serves as the control area operator for most of California and matches generation with loads. The CAISO starts with the CPX’s day-ahead bid schedules and with these schedules is responsible to maintain system reliability, coordinate generation dispatch, and provide open access to the grid. To fulfill these responsibilities, the CAISO can redispatch the CPX market results. There were three primary markets in California’s restructured electricity market under AB 1890: (1) the CPX Day-ahead market; (2) the CPX hourahead market; and (3) the CAISO Real-Time market. The CPX day-ahead market was designed and intended to be the primary wholesale electricity market. It worked in the following manner. Initially, supply and demand bids were submitted to the CPX 24 hours in advance. The CPX would then validate the bids and build supply and demand curves. The intersection of these curves established the single MCP received by all sellers. After the CPX market closed, those Scheduling Coordinators (SCs) whose bids were selected by the CPX would submit transmission access requests to the CAISO. The SCs could also submit offers to provide ancillary services1 to the CAISO. If the transmission requests caused congestion, the CAISO resolved these congestion problems by adjusting the SC’s requested transmission access and issuing usage charges on congested paths. The CPX also operated an hour-ahead market. In this market, participants could revise incrementally their schedules up to two hours before the trading hour begins. The third sequential market was the CAISO Real-Time market. This market was originally designed to accommodate about 3 percent of the total energy in the market and was designed to allow the CAISO to adjust load on a ten minute basis. The CAISO sorted the bids in price merit order and called upon the bids as necessary to balance generation and load. Each hour is divided into six ten-minute intervals, with separate incremental and decremental prices set during each interval. For two years, the CPX and CAISO markets operated more or less as designed, and wholesale prices (about $25 per MWH) were less than retail prices (about $50 per MWH) (or 2.5¢ per kWH versus 5¢ per kWH). During 2000, the relationship between the CPX’s day-ahead and the CAISO’s realtime market changed drastically with an unexpected shift to the CAISO’s real-time or energy-imbalance market began. In California’s wholesale power market, PG&E and SCE purchased nearly 90 percent of the energy traded. Due to the sequential nature of the 1

Spinning reserves, non-spinning reserves, regulation up and down, etc.

Design Flaws and a Worsening Crisis

65

market, with the CAISO’s real-time day-of market following the CPX’s dayahead market, buyers knew they had a second opportunity to purchase energy to cover their demand. This created an incentive for buyers to underschedule their next day’s electricity demand in order to reduce demand and drive down CPX’s single market-clearing price. By “underscheduling” in the CPX market, buyers hoped to pay less for their CPX purchases and would make up the remainder of their needs in the CAISO’s real-time market, paying higher prices for this portion of their needs. Sellers quickly responded to this strategy by underscheduling the supply side. By reducing the amount of energy available in the CPX’s day-ahead market, supply would decrease, potentially increasing the CPX’s marketclearing price. Sellers would sell fewer MWHs, but at a higher price in the CPX market. Sellers could then try to sell any remaining electricity in the CAISO’s secondary real-time market, perhaps at a lower price. This strategic bidding resulted in under-scheduling both supply and demand in the CPX market. This game resulted in a standoff in the CPX day-ahead market. However, the game caused the CAISO’s real-time market to grow in importance, ultimately increasing volatility and price levels. The IOUs had historically been able to forecast the next day’s electricity demand within about 2 to 3 percent of actual demand—that is, within 500 to 1000MW each day—depending upon the time of year. Underscheduling in 2000 significantly exceeded underscheduling in 1999. In 1999, the IOUs underscheduled by more than 2,000 MW a day only about a dozen times. In 2000, IOUs underscheduled by more than 2000 MW hundreds of times. During the last 40 days of 2000, underscheduling caused about 30 percent (not the 3 percent it was designed for) of the market’s MWHs to flow through the CAISO’s energy imbalance market. The CAISO would pay any price necessary to maintain system balance and reliability. Consequently this shift to the CAISO’s market caused wholesale price levels to increase drastically in California in 2000. The CAISO was often forced to make emergency OOM purchases at price and quantity levels that had been neither anticipated nor built into the initial market design. The CAISO was often hard pressed to locate the increased supply demanded in real time. This triggered power emergencies that signaled even higher prices. Often, the CAISO made OOM purchases at exorbitant prices to guarantee system reliability, which was the CAISO’s primary mandate. Subsequent to the time the FERC imposed price caps in California, these OOM purchases also became a means for sellers to avoid price caps. This may have been efficient. Regardless, the failure of California’s price caps to stem the crisis meant that the FERC had to adopt region-wide price caps that we describe below to tame the market in June 2001.

66

The California Electricity Crisis: What, Why, and What’s Next

Long-Term Contracts Were Not Available As illogical as it seems, the California competitive market was overregulated. The CPUC, FERC, and the CAISO all took shots at regulating this market. From this over-regulation came the requirement that virtually all IOU wholesale electricity be bought and sold in short-term commodity or spot markets. Forward and long-term bilateral contracts for electricity allow a purchaser to buy a certain amount of electricity at a pre-established price over some future period of time. These contracts provide future price certainty and supply security. However, even after the CPX began offering hedging instruments to the market, the CPUC initially restricted the value of these contracts to the IOUs by making their costs potentially unrecoverable under the continuing distribution company (DISCO) cost-of-service regulation. California’s reliance on spot markets was unique in restructured electricity markets. Table 7-1 shows that other states and countries restructured competitive electric markets under nearly opposite design conditions. Almost all power in these other restructured markets is sold under long-term or forward contracts. Very little power was sold in the spot market. Other deregulated energy markets sensibly used short-term commodity and spot markets to satisfy a small portion of their power requirements. These long-term contracts permitted these other jurisdictions to hedge that small percentage of power that was not secured through longterm contracts. California’s approach took the exact opposite route, exposing the majority of its market to the volatile spot market and severely limiting the ability of market participants to hedge this risk. Consequently, prices elsewhere were less volatile and single price MCPs were applicable to only a small fraction of the MWHs in those other markets.

Design Flaws and a Worsening Crisis

67

On May 26, 1999, FERC approved the CPX’s request to offer a block forward market service. The CPX’s new product was a long-term trading instrument designed to allow participants to hedge their short-term price risk. Buyers purchase 16-hour power blocks, from 6 a.m. to 10 p.m., for each day of the month except Sundays and holidays. The energy would then be delivered from one to six months following the month of the order. On February 24, 2000, FERC conditionally approved the CPX’s request to offer forward contracts to cover peak hours, when energy demand was highest. California’s two largest IOUs sought CPUC authority to buy future energy in the CPX block-forward market. In July 1999, the CPUC granted that authority. However, the CPUC limited the amount of power the two IOUs could buy through the forward market to one-third of their respective historical minimum hourly demand by month. Further, the utilities had to take delivery of these purchases no later than October 2000. The CPUC loosened these restrictions in March 2000, when it authorized requests by the two larger IOUs to increase their ability to forward contract through the CPX up to the amount of their respective “net short position.”2 The CPUC did, however, reserve the right to conduct future prudence or reasonableness reviews, subjecting the IOUs to potential refunds. The prospect for cost disallowance made the IOUs very wary about entering such long-term contracts. With no assurance that they would be allowed in the future to recover the contract costs from ratepayers, the IOUs did not fully utilize this hedging opportunity. As matters deteriorated in California, the CPUC began to ease these restrictions somewhat. The CPUC, in August 2000, allowed SCE and PG&E to enter into bilateral contracts ending by December 31, 2005. 2

3,000 MW for PG&E and 5,200 MWs for SCE.

68

The California Electricity Crisis: What, Why, and What’s Next

However, the CPUC did not expand the already established purchasing limits and again refused to guarantee cost recovery, leaving the IOUs subject to regulatory second-guessing in future prudence reviews. This essentially eliminated the IOUs’ incentives to enter into these long-term contracts. Consequently, the IOUs remained overly dependent on spot market sales during the summer of 2000. Market data show that the utilities did, in fact, underuse this hedging option. FERC and the CAISO data show that SCE used about 80 percent of its forward contracting authority of 2,200 MW in June 2000 and about 58 percent to 67 percent of its 5,200 MW limit for the months of July through August 2000. PG&E used approximately 37 percent of its 3,000 MW limit in June 2000 and roughly 60 percent of its limit in both July and August 2000. SDG&E did not participate in the CPX forwardcontract market at all during this period. Nevertheless, according to the CPX, even this limited use of the CPX forward market saved the IOUs about $706 million from May through September 2000. Not bad, but still too little and much too late.

The Retail Rate Freeze Between June and December 2000, conditions deteriorated rapidly. The IOUs were paying extremely high prices for their power and, because of AB 1890’s retail rate freeze, were unable to pass on to their customers these high prices. One cannot long buy high and sell low without incurring huge losses. The California IOUs were no different and soon incurred billions of dollars in liabilities. Credit rating firms watched the utilities’ deteriorating financial condition and, in January 2001, downgraded the California IOUs’ credit to junk-bond status. This meant that the IOUs could neither enter into long-term contracts nor purchase electricity from the spot markets. This forced the state of California, through the California Department of Water Resources (CDWR), to begin purchasing electricity for retail utility consumers in early 2001. Eventually, the state treasury’s financial credit ratings also dropped. The political and financial fallout continues into 2004. The FERC finally took action by issuing a mitigation order on December 15, 2000. This order eliminated the requirement that the IOUs buy and sell all electricity through the CPX. This action effectively eliminated the CPX as California’s primary wholesale spot market for electricity, leaving the state without a primary wholesale spot market. The CPX promptly went out of business and filed for bankruptcy in January 2001. The IOUs were thus free to enter into bilateral and forward contracts. Unfortunately, the IOUs’ creditworthiness, or lack thereof, meant that no generator was willing to enter into contracts with them. Consequently, starting mid-January 2001,

Design Flaws and a Worsening Crisis

69

the CDWR had to buy the power needed to meet utilities’ daily net short positions.

Price Caps and MWH Laundering The FERC addressed the higher prices in California by initiating a CAISO price cap of $250 per MWH during times of high demand. Likely, the price cap caused some sellers to bid into the CAISO market through OOM transactions, which were not subject to the price cap. This practice was known as “MWH laundering.” To see how this worked, assume that a $250 per MWH price cap is in effect at the CAISO. A generator could sell directly to a non-market participant (e.g., a municipal utility or out-of-state entity) for $350 per MWH. Non-market participants were not subject to the CAISO price cap, and could resell these same MWHs to the CAISO at a price unconstrained by the price cap (e.g., $375 per MWH) in an OOM trade. California is part of a larger western states energy market. This made it easier for participants to engage in MWH laundering. Thus, when FERC lowered the price cap in the CAISO, OOM purchases increased. We show this in Table 7-2.

Some California municipal utilities and out-of-state entities purchased electricity directly from generators and resold the electricity to the CAISO at prices at or above the cap. The FERC unwittingly accommodated this practice by converting its hard CAISO price cap into a so-called soft price cap under which OOM sales could exceed the price cap without affecting the single market-clearing price for in-state generators.

Scheduling Outages and Maintenance The CAISO had no authority over scheduled plant outages for maintenance. When the CAISO was being formed, it argued that it needed to be able to coordinate planned outages to effectively maintain the system’s reliability.

70

The California Electricity Crisis: What, Why, and What’s Next

The generation owners, however, contended that they should control scheduled outages. Ultimately, the generation owners prevailed. As a result, the CAISO had virtually no control over scheduled outages3 until the FERC took remedial steps in 2001 to correct this design flaw. The CAISO’s lack of authority to coordinate outages may have contributed to the problems in late 2000 when scheduled plant outages coincided with high demand. This decreased supply likely lead to higher prices. Unscheduled outages exacerbated these problems. Independent system operators for PJM, New York, and New England do have some control over scheduled outages. If the CAISO had similar authority in 2000/2001, it could have coordinated outages more effectively to help alleviate the price consequences of shortages in supply.

CONCLUSION Design flaws that had been masked when supply exceeded demand, were painfully exposed when market forces conspired to increase demand and reduce supply. Next, we explore several hypotheses to analyze econometrically these various facts and data.

3

The one exception was Reliability Must Run (RMR) contracts with certain generating units.

Chapter 8 TESTABLE HYPOTHESIS

The previous two chapters tell a story that mixes facts, data, and qualitative analyses. These may be sufficient for some. Others may not be so quick to accept the conclusions that come from this institutional approach to determine what went very wrong in California in 2000 and 2001. Therefore, in the next four chapters, we use an econometric approach to examine these same issues as testable hypotheses. This chapter identifies the hypotheses that we test. Chapter 9 reviews pricing and market power analyses conducted by others. Chapter 10 presents our natural gas price analyses. Chapter 11 presents our electricity price analyses. The general hypotheses are that market forces and structural designs in the California market affected natural gas and electricity prices. We concentrate on electricity prices in this book. Therefore, we examine and test the following hypotheses.

Market Forces World oil and natural gas price increases in 2000 and 2001 increased electricity prices in California. Droughts in the Northwest reduced imports to California and increased electricity prices in California. Extremely hot weather in 2000 increased demand and electricity prices in California. Unexpected economic population growth increased demand and prices in California.

72

The California Electricity Crisis: What, Why, and What’s Next NOx reclaim credit prices increased marginal running cost and electricity prices in California.

Natural gas prices in California were particularly influential after the surge in demand for electricity and the El Paso Pipeline explosion and increased electricity prices in California. We examined and tested each hypothesis. Applied econometricians should not be surprised by two difficulties we encountered. First, although natural gas and electricity data are available daily (and in the case of electricity hourly), other data we sought to examine either is not available on such a disaggregate basis or is not as meaningful on such a basis when it is available. Consider changes in economic activity. Daily changes in population, GDP, or employment are not available. These would likely be meaningless even if they were available. Daily rainfall is not meaningful when examining longer-term weather, such as droughts, even if it is available on such a disaggregate basis. However, cooling or heating degree day changes on a daily basis may be causally linked to daily price changes for electricity. Second, we previously described a “perfect storm” where several forces hit simultaneously. This tells a compelling institutional story and helps us understand all the things that went bad. Applied researchers see the same set of facts and correctly understand the difficulties of multicollinearity where several things turn bad, stay bad, and stop being bad on nearly the same days. In the analysis that follows to test our various hypotheses, we are particularly aware of and construct complex interdependent variables to reduce the multicollinearity problems. Third, we test natural gas hypotheses separately and test for the fact that natural gas and electricity prices are likely to be jointly determined variables. This means that some of the factors, such as the weather and petroleum markets that are presumed to affect electricity prices, might also affect natural gas prices. We treat gas prices as endogenous explanatory factors in our electricity price models Our primary hypotheses for natural gas are: Spot prices in California depend on prices at Henry Hub. Economic and climate factors affect natural gas prices. Storage affects natural gas prices. The El Paso Pipeline explosion increased the natural gas price in California.

Testable Hypothesis

73

Design Flaws Regulation and competition were mixed in California. There were emergency declarations, CPUC, and FERC policy changes. It would seem likely that these factors affected electricity prices in California. Within the limits of multicollinearity we tested the following hypotheses: Declared electricity emergencies caused California prices to increase, and these declarations may have been jointly determined with electricity prices. The FERC’s failure to implement a western-wide price cap, after MWH laundering and OOM purchases became known, caused California electricity prices to be higher. The CPUC’s failure to raise retail prices increased demand and wholesale electricity prices in California. The failure to use long-term forward contracts made electricity prices higher in California. Here, the problem of multicollinearity is particularly high. In the analysis discussed below, we combine several hypotheses into one. This is not completely satisfactory, but necessary. In the next three chapters, we discuss the specifications, data methods, and results of these hypotheses tests.

This page intentionally left blank

Chapter 9 SURVEY OF ELECTRICITY MODELS FOR CALIFORNIA

Various methods have been used to predict electricity prices in California. Both the CPX and CAISO developed models to predict market prices. Our review begins with an economic analysis of the CPX and CAISO approaches. These methods were developed before the energy crisis. Next, we discuss the econometric approaches that others have used in the various legal/regulatory proceedings that have addressed the question: “What caused the electricity crisis in California?”

THE CPX AND CAISO PRICING MODELS The CPX adopted a mean-reversion and a vector autoregression models to predict market-clearing prices. The principal authors were Seth Wilson, Robert Earle, and Karen Koyano.1 We discuss these models in turn. A Mean-Reversion Model generalizes the Geometric Brownian Motion model that has been used in Black-Scholes options theory.2 The basic conceptual model is:

1

2

See “Review of Price Behavior in the California Power Exchange,” Presentation to the CPX Board of Governors, May 18, 2000. F. Black and M. Scholes, “The Pricing of Options and Corporate Liability,” Journal of Political Economy, Vol. 81, 1973, pp. 637-659, and R.C. Merton, “Theory of Rational Option Pricing,” Bell Journal of Economics and Management Science, Vol. 4, Spring 1973, pp. 141-183.

76

The California Electricity Crisis: What, Why, and What’s Next

where P is the price, is a constant drift, measures price volatility, dP is the change in price in a small time period, dt is the time increment, and dw is a random Weiner process.3 The random term dw is assumed to be where is a standard normal distribution and has a zero mean and a variance equal to The Geometric Brownian Motion model assumes constant growth (per unit time) in prices. The Mean-Reversion Model modifies these generalizations and assumes that electricity prices revert to an equilibrium level and that in equilibrium there is no drift. Specifically, the Mean-Reversion Model restates equation (1) as:

where L denotes the equilibrium price level. When P exceeds L, there is a negative drift toward equilibrium, and when P falls short of L, there is positive drift. In discrete time, the Mean-Reversion Model is akin to the Koyck model for stock flow investments and has the form:4

The specific version of the model the CPX economists used to estimate prices introduced a stochastic term. The Mean-Reversion Model used by the CPX is a “mixing” model:

3

4

A continuous-time random process W(t) with such that W(t) – W(0) is normally distributed with mean 0 and variance t - s for and for which the increments are independent for non-overlapping time periods. I. Karatsas and S. Shreve, Brownian Motion and Stocastic Calculus, ed. (New York: Springer and Verlag, 1997); A. Papoulis, “Wiener-Levy Process,” Probability, Random Variables, and Stocastic Processes, ed.(New York: McGraw-Hill, 1984), pp. 292-293. L.M. Koyck, Distributed Lags and Investment Analysis (Amsterdam: North-Holland Publishing Company, 1954). See also M. Nerlove, “Distributed Lags and Demand Analysis,” U.S. Department of Agriculture Handbook 141, Washington, D.C., 1958.

Survey of Electricity Models for California

77

where dg is a second stochastic term. The additional term acts as a random jump that is specified to be independent of dw. Mean-reversion models are known in the finance literature as Vasicek models.5 Relatively simple formulas for the prices at European put and call options exist under mean-reversion and are discussed in Hull and White.6 The CPX specific model is estimated using a moving-average of past prices as an estimate of the equilibrium level L. Thus, the CPX model is completely identified by the historical time-pattern of prices. The CPX model is, consequently, not structural because it does not rely on the behavior or character of the economic market in which prices are set. Put simply, past prices predict current prices. The model does not explain why the equilibrium price level changes. Therefore, the model cannot detect why market prices surge or test the sensitivity of various economic factors. The deficiency in the mean-reversion model in explaining causal effects can be addressed by integrating it into a setting that explicitly determines structural effects. In this approach, the equilibrium price level L is specified to be a function of economic factors. This, in effect, is the approach we adopt in our econometric analyses of electricity prices in California. The second CPX model is known as a Vector Auto-Regression Model (VAR). The CPX’s VAR model estimates prices in the CPX’s day-ahead market (unconstrained market clearing prices or UMCPs) using the previous day’s price in the day-ahead market; the CAISO load forecast; natural gas prices for PG&E, SCE, and SDG&E city gates; temperature at San Francisco, Sacramento, Los Angeles, and San Diego; coal plant availability of the three IOUs; and nuclear availability of the three IOUs. The CPX VAR model was estimated using data for 731 days from April 1, 1998 through March 31, 2000. To avoid extreme multicollinearity, the explanatory variables are reduced to lower dimensional subsets using principal component analysis. In the PX VAR model, there are 58 parameters estimated: six each for load and squared load, two for unit availability, three for gas prices, eight for temperature, four interactions between gas prices and temperatures and 28 parameters for lags of UMCPs. The model has an auto-regression correction, which requires maximum likelihood estimation. 5

6

Vasicek, “An Equilibrium Characterization of the Term Structure,” Journal of Financial Economics, Vol. 5, 1977, pp. 177-188. John Hull and Alan White, “Pricing Interest-Rate-Derivative Securities,” The Review of Financial Studies, Vol. 3, 1990, pp. 573-592.

78

The California Electricity Crisis: What, Why, and What’s Next

The CPX used VAR models that were reduced-form and non-structural. Therefore, the CPX model mixed together supply and demand explanatory factors without any attempt to identify supply or demand. Commodity markets, such as the wholesale electricity markets in California, have buyers and sellers who hid in an auction. As we discussed earlier, there was a single price auction for the CPX. Market factors inform both sellers and buyers. The quantity demanded matters, but factors such as weather, economic conditions, and natural gas prices are generally more important because such information can inform ex ante bids. Quantity effects can be ambiguous. It is quite possible, and in California commonplace, for market clearing bid prices to be relatively high during low demand months, as well as during off-peak periods in low demand months (e.g., winter). Similarly, prices could be low during high demand months and during peak periods in the high demand months (e.g., summer). The CPX focused on market clearing price predictions and the underlying factors that may have influenced bidding behavior in the wholesale electricity market. Others that followed have generally followed this approach and, as a result, the quantity sold at any given point in time has generally been relegated to a secondary status. We now turn to the CAISO specific models. The CAISO approach employed engineering estimates of the incremental cost of supplying electricity. The concept behind this approach is that in competitive markets, sellers are price takers and would bid their short-run or incremental generating costs. As we explain below, the FERC adopted this conceptual approach for this reason in determining the Mitigated Market Clearing Price (MMCP), in the California Refund proceedings.7 The CAISO model simulates competitive market clearing prices using engineering estimates of incremental cost, which are then compared to actual market clearing prices. Wolfram (1998),8 Wolak and Patrick (2001),9 and Wolak (1999)10 adopted similar approaches. The CAISO approach adopts a competitive price benchmark standard that is estimated according to

7

San Diego Gas & Electric v. Sellers of Energy and Ancillary Service Into the Markets Operated by the California Independent System Operator and California Power Exchange, Docket Numbers EL00-95-075 and EL00-98-063. 8 C. Wolfram, “Strategic Bidding in a Multi-Unit Auction: An Empirical Analysis of Bids to Supply Electricity in England and Wales,” RAND Journal of Economics 29 (1998): pp. 703-725. 9 R. Patrick and F. Wolak, “Estimating the Customer-Level Demand for Electricity Under Real-Time Market Prices,” NBER Working Paper No. 8213, April 2001. 10 F. Wolak, “Market Design and Price Behavior in Restructured Electricity Markets: An International Comparison,” Stanford University Department of Economics.

Survey of Electricity Models for California

79

engineering supply curves measured at the point of residual demand. Borenstein, Bushnell, and Wolak (2000)11 applied a similar methodology. The objective of these approaches was to measure market power through the mark-up of price relative to marginal cost. This work was conducted for the CAISO’s Market Monitoring Committee, which defined market power in the Lerner sense as: “the ability of a firm, through its input or pricing decision, profitably to raise the market price above the competitive level.”12 The degree of market power is based on the gap between actual prices in comparison to engineering estimates of marginal costs. Similar pricing models discussed in the literature include Borenstein, Bushnell, and Knittel (1999)13 and Borenstein and Bushnell (1999)14. These both rely on a constant benchmark price using residual demand and engineering estimates of marginal cost. Other market power studies in the literature include Joskow and Kahn (2001),15 who also rely on a competitive price benchmark estimated by engineering methods, and Wolfram (1999), 16 who relies on a price-mark-up calculation with observed price and estimated marginal cost. This study differs from those previously described because it relies on the demand curve to calculate an elasticity adjusted price cost mark-up consistent with the Lerner mark-up. The CAISO-based models are not econometric models. They do not attempt to estimate electricity prices using econometric methods. Instead, the CAISO models used engineering estimates to derive marginal cost. These models estimate the degree of market power using price mark-ups above marginal cost. However, these market power models cannot distinguish the various causes of the differences between price and short-run marginal cost 11

S. Borenstein, J. Bushnell and F. Wolak, “Diagnosing Market Power in California’s Restructured Wholesale Electricity Market,” University of California Energy Institute, August 2000. 12 J. Bushnell, A. Klevorick, and R. Wilmouth. “Third Report on Market Issues in the California Power Exchange Energy Markets: The Impact of Reliable Must-Run Contract Reform and Ancillary-Services Market Redesign on the Performance of California’s Electricity Markets.” Prepared by the Market Monitoring Committee of the California Power Exchange on behalf of FERC, June 2000. 13 S. Borenstein, J. Bushnell, and C. Knittel. “Market Power in Electricity Markets: Beyond Concentration Measures,” University of California Energy Institute, February 1999. 14 S. Borenstein, and J. Bushnell, “An Empirical Analysis of the Potential for Market Power in California’s Electricity Industry,” Journal of Industrial Economics 47, September 1999, pp. 285-323. 15 Paul L. Joskow and Edward Kahn, “A Quantitative Analysis of Pricing Behavior in California’s Wholesale Electricity Market During Summer 2000,” AEI-Brookings Joint Center Working Paper No. 01 -1, January 2001. 16 C. Wolfram, “Measuring Duopoly Power in the British Electricity Spot Market,” American Economic Review 89, 1999, pp. 805-826.

80

The California Electricity Crisis: What, Why, and What’s Next

such as market power abuse, strategic bidding, or economic rent due to other market and structural problems.

OTHER ECONOMETRIC MODELS The California electricity crisis caused the FERC to consider refunds. The general premise of the various refund proceedings is that competitive markets are “just and reasonable” because sellers in competitive markets would bid their short-run marginal cost (SRMC). Consequently, economic rent would be paid to infra-marginal suppliers when the single MCP was greater than their short-run marginal cost of production. Consider Figure 91.

Sellers that operate between A and B have lower SRMC than the MCP. These sellers collect a mark-up over SRMC that pays down fixed and other risk-related components of their generation activity. The CAISO and FERC’s modeling approach, which focus on MMCP, are fully consistent with the proposition that MCPs for competitive markets would reflect SRMC. The difficulty that is masked in Figure 9-1 is that demand and supply curves shift in response to market forces and input prices. Therefore, various parties in the California refund proceedings have attempted to test various hypotheses to determine the factors that caused market prices to vary in California before, during, and after the refund period.

Survey of Electricity Models for California

81

Table 9-1 summarizes these various studies. The analyses of Harvey and Hogan are very similar to the analyses we perform. Daily electricity prices are analyzed and adjustments are made for serial correlation. The econometric analyses in Chapter 11 does both these things.

82

The California Electricity Crisis: What, Why, and What’s Next

Survey of Electricity Models for California

83

The CPX models constructed by Seth Wilson et al. were not prepared to test hypotheses concerning what caused fluctuations in California electricity prices. They do, however, support the idea that reduced form models with price as the dependent variable should be used to test hypotheses related to what factors affected the MCP in California. Miles Bidwell’s analysis is company specific and focuses on the degree of price mark-ups over marginal cost. It is a specific application of the CAISO modeling that was designed to test for the presence of market power. Mr. McCullough’s model is unique. He analyzes plant availability and outages and determines that 2000/2001 had excessive outages. He also has filed written testimony where he alleges that the significance of the drought in 2000/2001 is overstated. Mr. McCullough’s approach has had many critics. We will not repeat that here. Instead, we present his work in Table 91 for completeness. It has had little effect on what we did because our approach was concerned with prices, not plant availability. We next describe our econometric analysis of natural gas prices during the California electricity crisis.

This page intentionally left blank

Chapter 10 AN ECONOMIC ANALYSIS OF NATURAL GAS PRICE MOVEMENTS DURING THE CRISIS

Natural gas prices have a particular effect on electricity prices in California. This chapter extensively analyzes both the relationship between electricity and natural gas, and the prevailing market conditions in the natural gas industry during the energy crisis in California. We present statistical analyses that include daily natural gas price models for both Northern and Southern California markets for the period January 1, 1999 through July 31, 2002. These analyses test various hypotheses to determine both what caused movements in California’s spot natural gas prices and the relationship between California prices and national prices. It is undisputed that natural gas prices spiked during the electricity crisis in California. The facts behind this price spike are widely known and virtually undisputed. In October 2001, state analysts at the California Energy Commission (CEC) issued a report entitled Natural Gas Infrastructure Issues (CEC Report) that examines in detail the factors that caused the price spike in natural gas prices experienced in California in late 2000 and early 2001. Our structural analysis supports virtually all the CEC’s conclusions. We begin by summarizing basic facts. The natural gas market in California experienced its own “perfect storm” that reflects some of the same market-force factors that led to the electricity crisis in California. First, it is important to understand the market infrastructure. Five interstate pipelines deliver natural gas to California’s border. These pipelines have a total delivery capacity of about 7 billion cubic feet (Bcf) per day, and in-state production adds about 1 Bcf per day.1 The California intrastate delivery market has a Northern California market that is served by the PG&E intrastate pipeline delivery system, and the 1

CEC Report, p. 44.

86

The California Electricity Crisis: What, Why, and What’s Next

Southern California market is served primarily by the SoCalGas intrastate pipeline delivery system. The combined California intrastate pipeline system has a total capacity of 6.7 Bcf per day connected to interstate pipelines.2 About 15 percent of California’s natural gas is supplied from instate sources.3 The rest is sourced from Canada (28%), the Rockies (10%), and the Southwest (46%).4 In the year 2000, total demand in California was 6.584 Bcf per day, or 98 percent of California’s daily intrastate delivery capacity.5 While this shows a very tight statewide reserve margin, the actual reserve margin was even worse in Southern California, which was faced with severe supply shortages caused, in part, by demand exceeding capacity, and exacerbated by an inadequate infrastructure that limits receipts and constrains access to storage. While market conditions were tight for natural gas delivery, these markets remained reasonably workably competitive. Table 10-1 shows that interstate deliveries, plus in-state production, had a Hirschman-Herfindahl Index (HHI) of about 1897.

Table 10-2 shows that intrastate deliveries had an HHI of about 1503. HHIs of between 1000 and 1800 are generally deemed to be workably competitive.

2 3

CEC Report, p. 49. Ibid, p. 38.

4

California Energy Commission website report entitled “California Natural Gas Facts and Figures” at www.energy.ca.gov/naturalgas/natural_gas_facts.html.

5

California Energy Commission website report entitled “2000 California Natural Gas Consumption” at www.energy.ca.gov/naturalgas/consumption.html.

An Economic Analysis of Natural Gas Price Movements During the Crisis

87

As with the crisis in California’s electricity market, weather was a key market force affecting natural gas prices. Drought conditions in the Pacific Northwest reduced the hydroelectric supply to historically low levels.6 California found itself with very little excess capacity during peak demand periods.7 Recall that in California, natural gas-fired electric generators are the marginal power plants. With the decreased hydroelectric supply, these natural gas fired units had to be run more often to pick up the slack.8 This fact, coupled with the fact that many of these units were older and less efficient, contributed to a large increase in natural gas consumption. For example, between May and September of 2000, natural gas consumption by electric utilities in California consumed 22.4 percent more natural gas than

6

7 8

Hydropower from the Pacific Northwest declined from an hourly average of 20,805 MW in 1999 to 18,075 MW in 2000. California hydropower also declined from an hourly average of 4,395 MW in 1999 to 2,616 MW in 2000. See Cato Institute Report California’s Electricity Crisis: What’s Going on, Who’s to Blame, and What to Do, Jerry Taylor and Peter VanDoren (July 3, 2001) (Cato Report) citing Edward Krapels, “Was Gas to Blame? Exploring the Cause of California’s High Prices,” Public Utilities Fortnightly, January 1, 2000, p. 4. See also CEC Report, p. 74. Cato Report, p. 7. CEC Report, p. 74.

88

The California Electricity Crisis: What, Why, and What’s Next

during the corresponding months in 1999.9 In the West as a whole, electric generator demand for natural gas increased by 62 percent during this period.10 In California, electric generators’ natural gas consumption increased from 23% of total state consumption in 1999 to 35.3% in 2000, a 53.4% increase.11 This demand was even more pronounced in Southern California, where natural gas demand surged two to three times normal for the winter months.12 Recall that the Southern California natural gas intrastate pipeline and storage infrastructure was strained under normal situations. This increased demand further stretched the system. This tightening natural gas supply situation was exacerbated by an unseasonably hot summer in 2000. This hot weather increased natural gas consumption: between June 1999 and June 2000 by 7.3 percent in the Western Systems Coordinating Council states (excluding California) and by 13.7 percent for California. The CAISO’s average daily peak loads grew by 11 percent in May 2000 and 13 percent in June 2000 over corresponding periods for 1999.13 This unusually hot summer was followed by an extremely dry and cold winter throughout the West. This further taxed hydro conditions while customers increased their natural gas consumption to heat their homes.14 To make matters worse, there was an explosion in August 2000 on the El Paso Pipeline, which is one of the major pipelines delivering natural gas into Southern California. The pipeline rupture was not immediately rectified, and the El Paso Pipeline’s capacity into Southern California was still down 20 percent as late as October 2000.15 This had a dramatic effect on SoCal Gas’ storage and deliveries. Consequently, SoCal Gas entered the winter of 2000 with dangerously low storage levels.16 Although not nearly as serious as in Southern California, Northern California also experienced “slack” conditions in 2000-2001. In Northern California, capacity and supply barely

9

Cato Report, p. 8. 10

Ibid. See also S.A. Van Vactor and F.H. Pickle, “Money, Power, and Trade: What You Never Knew About the Western Energy Crisis,” Public Utilities Fortnightly, May 1, 2001, p. 36.

11

See California Energy Commission reports “California Natural Gas Facts and Figures,” www.energy.ca.gov/naturalgas/natural_gas_facts.html; “2000 California Natural Gas Consumption,” www.energy.ca.gov/naturalgas/consumption.html.

12

13 14 15 16

Between November 2000 and March 2001, SoCalGas’ winter electric generation natural gas demand ranged between .4 Bcf/d (November 2000) and 1.4 Bcf/d (January 2001) higher than the five year average. CEC Report Figure 5.4, p. 56. Cato Report, pp. 7-8. See also Van Vactor and Pickle, p. 36. CEC Report, p. 7. Ibid., p. 69. Ibid., p. 52.

An Economic Analysis of Natural Gas Price Movements During the Crisis

89

exceeded demand.17 In contrast, the Southern California natural gas infrastructure was simply swamped by market forces.18 To meet this demand, SoCalGas was forced to withdraw from its dwindling supply sources more than 1 Bcf/d to meet demand in January and February 2001.19 The notion of “slack capacity” is important when examining changes in natural gas prices. When there is excess capacity, competition will keep prices in line with the major North American trading centers, such as Henry Hub. When there is no excess capacity and regional demand outstrips capacity, prices will increase sharply in California and depart from Henry Hub. This concept can be readily observed. Since Northern and Southern California are served from different supply sources, price differentials can develop between Northern and Southern California due to differences in available pipeline capacity. In December 2000, both regional markets were tight due to the weather and a surge in worldwide oil prices. These two factors were continental in scope. Natural gas prices increased in Northern California to $14.58 per MMBtu.20 Southern California prices rose a bit higher, to $15.14 per MMBtu.21 During succeeding months, the regional price differential grew to as much as $5 per MMBtu22 due to a lack of pipeline supply in Southern California. The El Paso Pipeline explosion and a shortfall in natural gas storage contributed to this price gap. In 2002, however, demand, due to weakening economic conditions, dropped to levels substantially below demand in 2001. Consequently, natural gas supply and pipeline capacity again exceeded demand in both markets and price differentials all but disappeared.23 Supply and demand affects natural gas prices as they do all commodities. When transportation is unconstrained or supply exceeds demand, North American natural gas markets become highly interdependent and, in effect, move in lock step. Therefore, prices are highly correlated, and differences in price levels reflect transportation costs from Henry Hub back to producing areas and outward from producing areas that sell to more distant or more localized markets (e.g., intra-Alberta or in California). 17

CEC Report, pp. 55-56. According to the CEC Report, SoCal Gas operated at 101 percent of capacity in December 2000 and at 103 percent from January to March 2001 (pp. 55-56). 19 CEC Report, Figure 5.3 (p. 55). 20 Ibid, p.71. 21 Ibid. 22 Ibid. 23 Ibid. 18

90

The California Electricity Crisis: What, Why, and What’s Next

Just as the FERC recognized for electricity markets, there can be localized (fairly large, in fact) constrained markets for natural gas. Behind the constraints, market swings in prices are likely to be more volatile because these constrained natural gas markets respond to both worldwide petroleum conditions and localized swings in demand and supply. Some facts seem very important. These include the West Coast’s unique conditions during this period, including severe climate in late 2000 that combined drought and a cold late fall/early winter, insufficient summer storage fills, constrained pipeline capacity within the state (usage in excess of 100 percent utilization), and a phenomenal surge in electric system demand for natural gas. Nationwide, natural gas prices jumped dramatically from about $2.00 per MMBTU in 1999 to more than $10.00 per MMBTU at Henry Hub in 2000.24 These California-specific factors seem to have combined to compound or amplify the surge in prices. Southern California spot prices surged more than Northern California prices. In the ongoing Refund Case, the FERC recognized these facts by requiring a separate natural gas price for Northern and Southern California generation units, recognizing explicitly that there were two separate, independently operating natural gas markets in California. While both markets were severely affected by market forces that drove up natural gas prices in the state, Southern California was more severely affected due to the August 2000 pipeline explosion. The key point is that these two markets were both severely affected by a combination of market forces and infrastructure problems that were not faced by other markets in the country. There is no reason to think that natural gas prices would not spike in an area where demand severely outstrips supply. Nor is there any reason to think that prices in an area affected by unique market forces, which cause prices to spike, would be highly correlated to prices in an area not affected by those market forces. In fact, there is good reason to suspect just the opposite, that the prices in the two separate and diverse markets would not be highly correlated.

DAILY PRICE MOVEMENTS We now turn to an econometric analysis of daily natural gas price movements in Southern California and Northern California during the period from January 4, 1999 to July 31, 2002. Our analysis combines data from the Southern California market, where there have been allegations of supply manipulation (during late 2000 and early 2001), and the Northern California market where there were no such allegations. Importantly, our statistical 24

CEC Report, Figure 7.1, p. 71.

An Economic Analysis of Natural Gas Price Movements During the Crisis

91

analysis considers daily gas prices for time periods both before and after the alleged Southern California supply manipulation for each market in isolation. Our specification and selection of explanatory factors is similar to other studies published in the literature. For instance, Bopp (2000)25 analyzes daily price movements in natural gas. A related study is Walls (1994).26 The statistical models test for the statistical significance of: (1) daily Henry Hub index prices; (2) daily Gulf Coast fuel oil prices; (3) monthly natural gas consumption in California compared to normalized consumption; and (4) monthly California natural gas storage. The dependent variable for both Northern California and Southern California is the natural logarithm of the end-use market price for natural gas. We also used a dummy variable for winter because natural gas demand is seasonal. Our analysis also tested for the relevance (i.e., a statistical difference) of the “critical” eight-month time period from August 2000, when the explosion on the El Paso Pipeline occurred, through March 2001. This disruption in Southern California supply corresponds to the period FERC Administrative Law Judge Wagner found there was some evidence of pipeline supply shortages for Southern California.27 Table 10-3 shows the independent variables we used as explanatory variables in the regression equations. The daily natural gas spot prices we analyzed for Southern California (so_cal) and Henry Hub are the daily midpoint spot prices reported in Gas Daily. For Northern California, we used the average daily midpoint spot prices for Malin and PG&E Citygate reported in Gas Daily.

25

26

27

Anthony Bopp, “Daily Price Adjustments in the U.S. Market for Natural Gas,” Atlantic Economic Journal 28 (2000), pp. 254-265. D. Walls, “An Econometric Analysis of the Market for Natural Gas Futures,” The Energy Journal 16 (1995), pp. 71-84. 100 FERC ¶63,041 (September 23, 2002).

92

The California Electricity Crisis: What, Why, and What’s Next

From the outset, we found the data to be serially or auto-correlated. We used several functional forms and statistical estimation methods to correct this flaw because autocorrelation affects the validity of the hypothesis being tested. After experimenting with alternative models and specifications, we found specifications that included the lagged value of the daily spot price relative to the Henry Hub price (also lagged) effectively eliminated the serial correlation. The degree of auto-correlation apparent in the data is quite high. We conclude that the equations specified in Table 10-4 are the best statistical

An Economic Analysis of Natural Gas Price Movements During the Crisis

93

choice for Southern and Northern California daily spot price regressions during the time period January 1999 through July 2002. The final models were estimated using Ordinary Least Squares (OLS) regression.

The final specification of the independent variables that works best after various pre-testing is rather straightforward. First, we test the hypothesis that the daily price in the end-use market (e.g., Southern California) is correlated with the daily price at Henry Hub. This relationship was significant and positive in both the Southern and Northern California markets. In fact the

94

The California Electricity Crisis: What, Why, and What’s Next

elasticity of the pass-through from Henry Hub index prices to California prices was nearly one in both cases. Second, we tested the hypothesis that the current daily spot price in an end-use market also depends upon the previous day’s price differential between that market and Henry Hub. This econometric model uses a dynamic adjustment specification. Dynamic energy pricing models are discussed in Verleger (1982)28 for crude oil and Houthhakker, Verleger, and Sheehan (1974)29 for gasoline and electricity. In our approach, we include the lagged log price in comparison to the lagged price at the Henry Hub. To the extent that the lagged log price of gas is above the Henry Hub price, gas prices should adjust upward (i.e., the predicted price change is positive). This specification is also known as an error-correction or mean-revision model.30 In logarithmic form, the relevant variable is the logarithm of the ratio between yesterday’s Southern California spot price and yesterday’s Henry Hub spot price. This hypothesis holds with significant probability for both Southern and Northern California. Next, we tested the daily Gulf Coast Heating Oil price as an explanatory factor because it is known to move seasonally with natural gas prices and because it reflects worldwide crude oil price movements over time. (In other models, we found that Gulf Coast Heating Oil prices generally outperform World Crude Oil prices in natural gas regression equations.) That said, one can generally only accept the hypothesis of Gulf Coast Heating Oil prices being positively correlated with daily spot natural gas prices marginally, if at all. We considered various natural gas consumption measures as possible explanatory factors affecting natural gas price movements. The hypothesis is that when electricity demand, economic factors, or climate conditions push up demand for natural gas, spot prices would increase, especially if supply is tight. There are several dimensions to such a hypothesis. First, we used normalized monthly natural gas consumption (see Table 10-3). A very cold month or a period with a high level of economic activity would mean that normalized monthly natural gas consumption (demand) would be higher. Since this would be especially true in the critical winter period, we 28

Philip K. Verleger, “The Determinants of Official OPEC Crude Oil Prices,” Review of Economics and Statistics 64 (May 1982), pp. 177-183.

29

H.S. Houthhakker, Philip Verleger and Dennis Sheehan, “Dynamic Demand Analysis for Gasoline and Residential Electricity,” American Journal of Agricultural Economics (May 1974), pp. 412-418.

30

For an introduction to error-correction model, see W. Greene, Econometric Analysis. Ed., Chapter 17. The error correction specification we use is: where is the logarithm of gas price (Northern or Southern CA), is the logarithm of the Henry Hub gas index price, are additional explanatory factors with weights and is the unobserved error term.

An Economic Analysis of Natural Gas Price Movements During the Crisis

95

separately considered the normalized gas consumption measure for the post pipeline explosion period during the winter (CRISIS). We expected natural gas supplies to be tighter than normal in Southern California and, to some lesser degree, statewide in these instances. Two variables emerged from this analysis: (1) Normalized Consumption and (2) Normalized Consumption During Crises. Contrary to our expectations, neither variable was significant in either the Southern California or Northern California markets. Next, we test for the significance of monthly storage in California. Storage effects on natural gas prices are discussed in Susmel and Thompson (1997)31 and more generally in Deaton and Laroque (1992,32 199633) and Wright and Williams (1982)34. When monthly storage declines, daily spot prices increase in both California markets. This result was confirmed in both California markets. Finally, the winter indicator (WINTER) showed that prices were marginally lower in the winter, all other factors held constant, although the seasonal price pattern was different in the two California markets. We have already discussed some tests for the structural stability of these results over time. We also considered whether the relationship of natural gas prices in California markets to the Henry Hub market was significantly different during the crisis period. In the Southern California market, the coefficient on the Henry Hub variable in the critical period was negative, suggesting that the partial correlation of Southern California prices to the Henry Hub pass-through elasticity was smaller in the critical period than otherwise. This result did not, however, reach statistical significance. The Northern California market revealed no significant change in the relationship of Northern California natural gas prices to the Henry Hub during the critical period. More interestingly, we found that natural gas prices in Southern California rose after the pipeline explosion, but that Northern California prices were not similarly affected. Figures 10-1 and 10-2 show the actual and predicted daily spot prices for Southern and Northern California using the OLS equations in Table 10-4.

31

Raul Susmel and Andrew Thompson, “Volatility, Storage and Convenience: Evidence from Natural Gas Markets,” The Journal of Futures Markets 17 (1997), pp. 17-43.

32

Angus Deaton and Guy Laroque, “On the Behavior of Commodity Prices,” Review of Economic Studies 59 (1992), pp. 1-23.

33

Angus Deaton and Guy Laroque, “Competitive Storage and Commodity Price Dynamics,” Journal of Political Economy 104 (1996), pp. 896-923.

34

Jeffrey C. Williams and Brian D. Wright, “The Economic Role of Commodity Storage,” The Economic Journal 92 (Sept 1982), pp. 596-614.

96

The California Electricity Crisis: What, Why, and What’s Next

There is very little difference between actual and predicted natural gas prices during the roughly three and one-half year period.

Next, these same comparisons are shown in Figures 10-3 and 10-4 for the period January 1, 2000 through June 20, 2001 to correspond to the expanded period the FERC has permitted for scrutiny in the Refund Case, which we will discuss in greater detail in Chapter 11. Again, there are only minor

An Economic Analysis of Natural Gas Price Movements During the Crisis

97

differences between actual and predicted values using these regression equations.

98

The California Electricity Crisis: What, Why, and What’s Next

SIMULATION ANALYSIS Regulatory attention focused on the natural gas supply disruption in the eight months subsequent to the El Paso Pipeline accident. Thus, the focus is on the predicted price differences in the eight critical months subsequent to the El Paso Pipeline explosion relative to the other months in the analysis. Here, the four winter months (December, January, February, and March) are most important. For the critical period, one can predict what prices would have been if all explanatory factors had been the same as they were in actuality, except that the period of time was normal (i.e., the period of time where the dummy variable for the critical period would be equal to zero). In this “but for” analysis, we predict natural gas prices assuming that the time periods were neither critical nor in crisis. In performing this simulation, we eliminated variables appearing as interactions with the critical period dummy if they had insignificant coefficients. Since no critical period variables were significant in the Northern California regressions, we do not present a “but for” analysis in this case. Table 10-5 presents the modified regression models used in the simulations.

An Economic Analysis of Natural Gas Price Movements During the Crisis

99

We also used the modified regression model to predict daily Southern California natural gas prices during the fourteen-month period from May 2000 through June 2001. These prices can be compared to the actual prices during this period. These are shown using monthly averages in Table 10-6 for Southern California and in Table 10-7 for Northern California.

100

The California Electricity Crisis: What, Why, and What’s Next

For Southern California, the difference between predicted and actual prices increases in the crisis months when we exclude the critical period effect. With respect to the regulatory scrutiny period (after the pipeline

An Economic Analysis of Natural Gas Price Movements During the Crisis

101

explosion), actual prices averaged $12.08 for Southern California. The best prediction is that prices would average $11.93, or 15¢ less. This difference is about 1.2 percent of the actual price. Without adjusting for the critical period factors, the predicted price would decrease by 27¢ from the predicted price. Hence, we conclude that the pipeline explosion caused an estimated 2.3 percent price increase in Southern California natural gas prices during the critical period. During the four-month crisis period, the average predicted price was $17.56. Without adjusting for critical period factors, the average predicted price was $17.00. This is a 56¢ decrease from the predicted actual price. Hence, we conclude that the El Paso explosion caused prices to increase in Southern California by 3.2 percent during the crisis period. In Northern California, the effect of the El Paso Pipeline explosion on natural gas prices was not significant. The predicted price for the crisis period averaged $9.06, 14¢ or 1.5 percent lower than the actual price. Since the critical period effects were insignificant in Northern California markets, the predicted price differentials demonstrating the effects of the El Paso Pipeline explosion are equal to zero.

CONCLUSION Our main conclusions are that underlying factors, especially the prices at and relative to Henry Hub, explain the underlying natural gas prices to a considerable extent. The daily natural gas analysis rejects the hypothesis that natural gas prices were somehow artificially inflated during this period. If the natural gas markets had been manipulated, we would expect to see evidence of higher prices in both Northern and Southern California, all other things equal. This was not the case. Evidence shows that Southern California prices were higher due to the El Paso pipeline explosion. The alternative hypothesis that the critical period was a period of market manipulation must be rejected unless it is plausible that only Southern California prices were subject to manipulation. Finally, the relationship of California prices to the Henry Hub was not dramatically shifted in the period that subsequently received regulatory scrutiny. Yet, this relationship with Henry Hub prices does not, by itself, completely explain price setting in California. Other factors, including the limited gas storage, excess demand, seasonality, and the pipeline explosion, help explain California prices. Thus, a formula for setting prices during the refund period based solely on Henry Hub index prices and transportation costs must certainly miss the mark as it cannot

102

The California Electricity Crisis: What, Why, and What’s Next

possibly adjust for the pipeline explosion, limited gas storage, or the excess demand for natural gas.

Chapter 11 AN ECONOMETRIC ANALYSIS OF ELECTRICITY PRICES IN CALIFORNIA

This chapter analyzes daily electricity spot prices in western United States markets. In part, this analysis uses the results for natural gas discussed in Chapter 10. We test various hypotheses discussed in Chapter 8. In addition to the models discussed in detail in Chapter 9, several studies have also analyzed spot electricity prices. The majority of published studies concern the relationship between spot and forward electricity prices. Studies that compare forward and spot prices include Shawky, Marathe, and Barrett (2003),1 Woo, Horowitz, and Hoang (2001)2 and Amundsen and Singh (1992).3 Pricing and volatility is discussed in Barlow (2002),4 Robinson and Baniak (2002)5 and Green (1996).6 Optimal hedging is discussed in Wolak (2001)7 and Bessembinder and Lemmon (2002).8 Our approach follows the 1

Hany A. Shawky, Achla Marathe, and Christopher L. Barrett, “A First Look at the Empirical Relation between Spot and Futures Electricity Prices in the United States,” Journal of Futures Markets 23 (Oct. 2003), pp. 931-955. 2 Chi-Keung Woo, Ira Horowitz, and Khoa Hoang, “Cross Hedging and Forward-Contract Pricing of Electricity.” Energy Economics 23 (Jan. 2001), pp. 1-15. 3 Elrik Schroder Amundsen and Balbir Singh, “Developing Futures Markets for Electricity in Europe,” Energy Journal 13 (1992), pp. 95-112. 4 M.T. Barlow, “A Diffusion Model for Electricity Prices,” Mathematical Finance 12 (Oct. 2002), pp. 287-298. 5 Terry Robinson and Andrzej Baniak, “The Volatility of Prices in the English and Welsh Electricity Pool,” Applied Economics 34 (Aug. 2002), pp. 1487-1495. 6 Richard J. Green, “Increasing Competition in the British Electricity Spot Market,” Journal of Industrial Economics 44 (June 1996), pp. 39-52. 7 Frank Wolak, “An Empirical Analysis of the Impact of Hedge Contracts on Bidding Behavior in a Competitive Electricity Market,” National Bureau of Economic Research Working Paper 8212 (Apr. 2001). 8 Hendrik Bessembinder and Michael Lemmon, “Equilibrium Pricing and Optimal Hedging in Electricity Forward Markets,” Journal of Finance 57 (June 2002).

104

The California Electricity Crisis: What, Why, and What’s Next

practice of others who analyzed spot prices based on reduced form models. However, we complicate matters by testing for the presence of jointly determined endogenous variables. Specifically, this analysis focuses on the daily spot price movements in several electricity markets, including both western trading hubs and California markets. We distinguish between peak and off-peak markets, while eliminating non-trading days for these spot markets. We consider several explanatory factors to test specific hypotheses related to world oil prices, natural gas prices, weather, power emergencies, available supply, demand, economic conditions, and market and regulatory design flaws. There are some anomalies in the data that require explanation. Both peak and off-peak electricity price data are available for the western states’ spot markets. These are the Dow Jones spot markets, which consist of California-Oregon Border (COB), Mid-Columbia (Mid-C), Four Corners, Palo Verde, SP15, and NP15. The Dow Jones SP15 and NP15 spot markets were not formed until December 2000. Prior to that time, spot electricity was sold in California through the CPX. We analyzed the unadjusted hourly prices for the CPX data for the period January 1999 to January 2001 when the CPX spot markets were open. By contrast, the CPX markets are statewide, not regional. Accordingly, a statewide average natural gas price is relevant for the CPX analysis. Finally, there are several days on which there is zero off-peak volume reported for Dow Jones spot sales. We dropped these observations from this analysis. We divide this statistical analysis of daily electricity prices into two segments: (1) California (NP15, SP15, and PX), and (2) western states (MidC, COB, Four Corners, and Palo Verde). Table 11-1 shows the variables we found best explained electricity price levels in California and therefore, formed the hypotheses rejection tests of statistical reliability in this analysis.

An Econometric Analysis of Electricity Prices in California

105

106

The California Electricity Crisis: What, Why, and What’s Next

Table 11-2 lists the dependent price variables we used in this analysis. The model specification shown is logarithmic. We tested both arithmetic and logarithmic forms. There was no discernable difference in the hypotheses testing or reliability results. There was a slight improvement in or goodness of fit. Therefore, we present the logarithmic form here.

An Econometric Analysis of Electricity Prices in California

107

The model of spot electricity prices in California and western markets is heavily dependent on regional natural gas prices. Many of the same factors that reasonably could have affected daily electricity prices in western markets could also have affected daily natural gas prices in the west. In addition, natural gas prices affect the cost of producing electricity. The marginal electricity supplier in California often relies on natural gas generation for in-state supply. These considerations imply that daily electricity prices and natural gas prices are jointly determined. In the analysis that follows, we treat natural gas prices as endogenous factors in the electricity price equations. In our procedure to achieve consistent estimates of the regressions parameters, the actual daily natural gas price may be replaced with a predicted variable that is not jointly determined. The prediction step requires additional instrumental variables that affect the natural gas price, but do not directly influence electricity prices. These instruments provide the necessary econometric identification. A statistical test of simultaneity based on the statistical significance of the predicted gas price coefficient was performed in this regression. Due to serial correlation, we use a generalized least squares (GLS) approach for the regression equation with both natural gas and predicted natural gas prices used as explanatory factors to produce consistent estimates of parameters that are equivalent to the instrumental variable estimates.

108

The California Electricity Crisis: What, Why, and What’s Next

The model specification includes factors for the maximum system load and the traded volume in the relevant regional market. Given the short-run inelasticity of demand in the spot market, we treat volume exogenously. This is consistent with the work of others discussed in Chapter 9. Other exogenous variables include the California unemployment rate (representing the state of the economy), the ratio of combined nuclear output and hydroelectric generation in California relative to imports of electricity from Canada (representing supply availability), monthly dummies (to capture seasonality in demand and pricing), and several period and event dummies. The refund period dummy determines the amount by which electricity prices increased during the refund period compared to earlier and later periods. The regulatory adjustment dummy indicates a time period where several regulatory and market design flaws were identified and corrected. The hypothesis tested is that while prices were higher in some western markets during the refund period, the change in regulatory policy and correction of market flaws reduced California electricity prices substantially. Finally, we include a factor for the declared emergencies by the CAISO as a potential signal of extreme market under-capacity. On days with declared emergencies, we hypothesize that prices would be greater for two reasons. First, prices may rise because of scarcity when there is increased demand and decreased supply. Secondly, prices potentially rise if market participants are able to alter trading strategies as a result of the information signal inherent in the declared emergency. Reasonably, declaring an emergency is itself an endogenous event caused by market conditions related to supply and demand. To treat this factor endogenously, we first perform a logit9 analysis of declared emergencies. The logit analysis allows us to determine the probability that an event occurs based on observed explanatory factors. We then use a selection correction technique (Dubin and McFadden (1984))10 to add an additional factor to the regression model that accounts for the potential correlation of the event dummy and the error term of the electricity price equation. The addition of the selection correction term permits consistent estimates of the parameters 9

The logit model has the form: where X is a vector of factors explaining the discrete choice and are associated weights. Maximum likelihood is used to fit the parameter given observed choices and corresponding explanatory factors. For additional details, see Daniel L. McFadden, “Conditional Logit Analysis of Qualitative Choice Behavior,” Frontiers of Econometrics (1973), New York: Academic Press. 10 The selection correction term is: where ev_days is a dummy variable for declared emergencies and P is the predicted probability of the declared emergency from the logit model. For additional details, see Jeffrey A. Dubin, and Daniel L. McFadden, “An Econometric Analysis of Residential Electric Appliance Holdings and Consumption,” Econometrica 52 (March 1984), pp. 345362.

An Econometric Analysis of Electricity Prices in California

109

for both OLS and GLS and also allows us to determine whether the event dummy is statistically endogenous. Table 11-3 shows the logit analysis of the declared event emergency.

110

The California Electricity Crisis: What, Why, and What’s Next

This analysis shows that emergency declarations depend on weatherrelated factors that shifted demand (cooling degree days) and supply (rainfall in the Northwest), plus other seasonal factors (monthly dummies). Specifically, we use several climate measures to predict declared emergencies. In particular, we used two factors to measure rainfall

An Econometric Analysis of Electricity Prices in California

111

(measured as average daily levels and monthly levels in Oregon and Washington). We used an additional factor to reflect air conditioning needs in terms of California cooling degree days. These factors are not included in the spot electricity pricing analysis (for identification) and are reasonably associated with climatic conditions that may lead to declared emergencies. There were 894 daily observations analyzed (after combining peak and offpeak markets). We observed 110 declared events, or 12 percent. The model correctly predicted 88 percent of the observations using logit estimation. The climatic or weather effects were statistically significant determinants of declared emergencies. Specifically, greater levels of cooling degree days (hotter climatic conditions) raise the probability of an event. Greater levels of rain (leading to greater hydroelectric supply) were negatively associated with declared events. There was some discernable seasonality to declared events, as might be expected.

CALIFORNIA TRADING MARKETS We test the various hypotheses we discussed in Chapter 8 in our regression analysis for the California spot electricity markets. This estimation and testing used GLS to correct for first order serial correlation. An additional complication is the presence of jointly determined, or endogenous variables, as explanatory variables. The procedure we used follows Hatanaka (1976),11 Fair (1970),12 and Fair (1972).13 Table 11-4 (column 1) shows the statistical analysis of the Dow Jones’ spot electricity market for peak hours over the 402 weekdays (mid-December 2000 through July 31, 2002) during which electricity was traded in NP15.

11

12

13

Michio Hatanaka, “Several Efficient Two-Step Estimators for the Dynamic Simultaneous Equations Model with Autoregressive Disturbances,” Journal of Econometrics 4 (May 1976), pp. 189-204. Ray C. Fair, “The Estimation of Simultaneous Equation Models with Lagged Endogenous Variables and First Order Serially Correlated Errors,” Econometrica 38 (May 1970), pp. 507-516. Ray C. Fair, “Efficient Estimation of Simultaneous Equations with Auto-Regressive Errors by Instrumental Variables,” Review of Economics and Statistics 54 (Nov. 1972), pp. 444449.

112

The California Electricity Crisis: What, Why, and What’s Next

About 97 percent of the variance in the logarithm of these prices is explained by the regression equation. During this year-and-one-half period, peak spot electricity prices in California increase when natural gas prices

An Econometric Analysis of Electricity Prices in California

113

increase. Prices were also higher during the refund period. A declared event emergency has a positive, but insignificant, effect on prices. The maximum daily load effect was significant and had a positive effect on spot price in this equation, but the volume-traded effect was not significant. Peak spot electricity prices decline when in-state supplies increase and imports decline. Peak prices also move counter to the state’s unemployment rate: increasing when the economy grows and falling when the economy declines. We also find that prices decreased when regulatory and market design flaws were corrected during the refund period. The predicted level of natural gas is significant in the equation.14 Table 11-4 (column 2) shows the corresponding statistical analysis for Dow Jones’ off-peak spot prices during this same period. About 96 percent of the variance in prices is explained in the equation. The maximum daily load and refund period variables both have insignificant coefficients, while trading volume has a significantly positive coefficient. Other variables have similar effects on off-peak spot electricity prices as compared to the peak market during this time period in Northern California. Table 11-4 (columns 3 and 4) shows the corresponding statistical results for SP15. About 97 percent of the total peak price variance is explained in the equation. The same explanatory variables found in the SP15 peak model are also present in the NP15 peak model. The predicted natural gas price variable differs in the two models because a corresponding regional Northern and Southern California spot natural gas price is used in the respective regional electricity equations for NP15 and SP15. A second difference is that the volume of SP15 peak electricity sold is statistically significant. Volume sold increases as prices decline, and vice versa. Otherwise, the results are quite similar in columns 1 and 3. There is a slight weakening in the relevance of the regulatory and market design correction factor for SP15 relative to NP15. Column 4 shows the corresponding Dow Jones spot off-peak results for this period. About 95 percent of the variance in price is explained in the equation. The off-peak equation for SP15 is similar to the corresponding peak equation. The relative supply availability factor was not significant in this market. Columns 5 and 6 are California statewide, not regional, spot market regressions. We derived the spot price data from the CPX’s unconstrained spot hourly prices. The time period is non-holiday weekdays during the period January 1999 through January 2001. Since these equations are statewide, an average predicted statewide natural gas price was used rather than predicted regional natural gas prices. About 93 percent of the daily 14

World oil prices were not significant determinants of western California spot electricity prices in this or the other models and were eliminated from the equations.

114

The California Electricity Crisis: What, Why, and What’s Next

price variance is explained in the equation. The model for peak-period spot prices is shown in column 5. Higher daily natural gas prices and max daily load caused spot electricity prices to increase in the CPX markets while the correction of CPUC and long-term contract flaws caused daily electricity prices to decrease in the CPX markets during the time period from January 1999 through January 2001. Column 6 shows the corresponding spot offpeak results for the CPX markets during this January 1999 through midJanuary 2001 time period. The statistical results are similar to those shown in column 5. The coefficient on the refund period dummy in the CPX regressions was not significant and, thus, does not demonstrate increased spot prices in the CPX markets during the refund period. The figures below compare the actual and predicted prices from the California markets.

An Econometric Analysis of Electricity Prices in California

115

116

The California Electricity Crisis: What, Why, and What’s Next

An Econometric Analysis of Electricity Prices in California

117

WESTERN TRADING MARKETS The data for the western trading markets comes from a single source. About 900 daily observations are available for spot peak sales and most off-peak sales from January 1, 1999 through July 31, 2002. Table 11-5 shows the four western trading markets and the variables that were statistically significant. These market regions are: (1) Mid-C; (2) COB; (3) Four Corners; and (4) Palo Verde.

118

The California Electricity Crisis: What, Why, and What’s Next

An Econometric Analysis of Electricity Prices in California

119

The first two are northern markets and we use a northern natural gas price to test the hypothesis that changes in predicted natural gas prices cause changes in daily western spot electricity prices. The other two western markets are southern. We use southern predicted spot natural gas prices as explanatory factors in these markets. Table 11-5 (column 1) shows the factors that combine to explain 92 percent of the daily changes in Mid-Columbia spot peak electricity prices. Similar explanatory variables were found to matter for spot-market peak electricity prices in California and Mid-Columbia. Higher predicted natural gas prices, max daily load, and the refund period lead to higher prices, while correcting the regulatory and market design flaws reversed this increase to a degree. Lower monthly unemployment in California pushes up demand and daily Mid-Columbia peak electricity prices. Table 11-5 (column 2) shows the off-peak Mid-Columbia regression analysis that explains about 92 percent of the variance in the spot off-peak electricity prices. The same variables that were found to explain MidColumbia peak prices are also significant in the off-peak market. One exception is the statistical magnitude of the max load effect, which is lower in the off-peak equation than in the peak equation. This result is found in other markets and is reasonably expected because it is less likely for offpeak demand, in comparison to peak demand, to be affected by factors that push up maximum demand. Table 11-5 (column 3) shows the regression results for spot peak electricity prices at COB. The regression explains about 94 percent of the variance in daily electricity prices. The same variables that caused changes in Mid-Columbia peak prices are also statistically significant for COB peak prices. A noteworthy aspect of this model is the strong inverse relationship between volume sold and the spot price in the COB market for spot peak sales. The regression for daily off-peak spot electricity at COB explains about 93 percent of the daily variance in off-peak electricity prices at COB. The same variables and patterns found to be significant in the COB off-peak equation were also found to be significant in the COB peak equation. Table 11-5 (columns 5 through 8) is the southern western market counterpart to Mid-Columbia and COB. Column 5 shows the regression results for Four Corners daily spot peak electricity prices. The regression explains about 94 percent of the daily price variance. Column 6 shows the off-peak Four Corners daily spot price regression results. The regression explains about 92 percent of the daily price variation. Column 7 shows the regression results for daily spot peak electricity prices at Palo Verde. The regression explains about 94 percent of the variance in prices during the period. Generally, the relative levels of California internal to external

120

The California Electricity Crisis: What, Why, and What’s Next

electricity supply sources were not significant in the western market regressions. This result should be expected. The figures below display the actual and predicted daily electricity prices for western markets.

An Econometric Analysis of Electricity Prices in California

121

122

The California Electricity Crisis: What, Why, and What’s Next

An Econometric Analysis of Electricity Prices in California

123

124

The California Electricity Crisis: What, Why, and What’s Next

ADDITIONAL TESTING Statistical tests of endogeneity were also performed (Hausman (1978)).15 If the t-statistic of the predicted natural gas price variable is significant in a regression with both natural gas price and predicted natural gas price, then the corresponding electricity and natural gas price variables are simultaneously determined. In all equations, we find endogeneity of natural gas prices in the spot electricity equations based on the Hausman Test. This means that the adjustment in the electricity equations discussed above is necessary to achieve consistent statistical results. The event dummy was generally positive, but not statistically significant in the western markets. The event dummy variable was also positive, but only marginally significant in the California markets. The selection correction term did not indicate endogeneity of the declared emergency factor. The refund period dummy supported the hypothesis of higher prices in all markets other than the CPX markets. For example, in the NP15 market, the refund period dummy showed that prices were higher by 146 percent.16 After regulatory and market design flaws were corrected, prices fell to 90 percent of non-refund levels.17 Prices in the refund period were also higher in the western markets. For instance, the market for Mid-Columbia peak electricity experienced 162 percent higher prices in the refund period. 18 Correcting market and regulatory flaws led to price levels that were subsequently 119 percent higher than non-refund period prices.19

CONCLUSION These analyses confirm that virtually all variations in pricing during the period January 1, 2000 through June 20, 2001 can be explained by fundamental economic forces and specific regulatory policies in these 15

See Hausman, J. “Specification Tests in Econometrics,” Econometrica 46 (1978): 12511271. 16 The coefficient of the refund period dummy in the peak NP15 model was 0.376. exp(0.376) = 1.46. 17 The coefficient of the corrected flaw dummy in the peak NP15 model was 0.478. exp(0.376 – 0.478) = 0.90. 18 The coefficient of the refund period dummy in the peak Mid-C model was 0.483. exp(0.483) = 1.62. 19 The coefficient of the corrected flaw dummy in the peak Mid-C model was 0.306. exp(0.483 – 0.306) = 1.19.

An Econometric Analysis of Electricity Prices in California

125

markets. The markets created by the CPUC, the California Legislature, the CPX, then-Governor Davis, and the FERC profoundly influenced some of those forces. The reversal in prices after the market and regulatory flaws were eliminated is as important as the demonstrable price increases. For some, this econometric analysis may be complementary to the institutional discussion in Chapters 6 and 7, confirming their belief that the FERC refund and market manipulation proceedings may be misguided or at least excessive. Before addressing such matters, Chapter 12 introduces a new major concern: market manipulation. Here the question is not whether spot price movements can be predicted. The fundamental question is whether some market participants may have acted within or outside the market rules to manipulate the market to such an extent that the market clearing prices were not reasonably competitive.

This page intentionally left blank

Chapter 12 MARKET MANIPULATION

Our previous discussion supports the conclusion that the California electricity crisis ran deep and that many factors contributed to this unprecedented breakdown of markets and politics. In summary: There were significant, mostly unexpected, market forces that increased demand and shorted supply throughout the western United States. World petroleum prices, especially natural gas, jumped wildly in 2000 and 2001. There were fundamental design flaws in California’s restructured wholesale electric power markets. The California market restructuring flourished initially because excess supply masked market design flaws and market forces were favorable. However, when the market forces shifted in 2000, these design flaws were exposed, and, some allege, were exploited by certain market participants. In this chapter, we review those allegations, how market monitoring and detection of wrongdoing worked, and the FERC’s response to these market manipulation allegations. Commodity markets have winners and losers. Some participants trade and speculate. Others provide hedging and insurance services to market participants. Commodity markets are often volatile, particularly those that focus on short-term or spot transactions. California’s electricity markets had all of the above. In the aftermath of the crisis, many have engaged in an ex post “blame game.” In this game, winners are presumed to have behaved badly and were not just “smart” or “lucky.” Some significant wrongdoing has been detected. The investigations and inquiries continue into 2004.

128

The California Electricity Crisis: What, Why, and What’s Next

ALLEGATIONS OF UNJUST AND ANOMALOUS TRADING Several different reports, memos, and orders have identified potentially problematic or anomalous trading strategies. The first was the nowinfamous Hall/Yoder memo dated December 6, 2000, which discussed several of Enron’s trading strategies. This in turn led to a FERC Staff investigation and an initial report (August 2, 2002) in a fact-finding investigation into the Enron trading strategies (Docket No. PA02-2-000). The FERC Staff subsequently published its Final Report on Price Manipulation in Western Markets in that docket on March 2003. The California Parties1 filed lengthy testimony on February 25, 2003 in the California Refund Proceedings (Docket No. EL-00-95-000) where they identified various gaming strategies and identified the entities the California Parties allege may have been involved in these strategies. On June 2003, the CAISO released its report on alleged gaming in the California market, identifying those companies it alleges were guilty of market manipulation and game playing in violation of the CAISO tariff. Finally, in a series of Show Cause Orders, the FERC began a company specific investigation with respect to several of these allegations. Below, we review the various allegations that have been made, starting with the activities identified in the Enron memo written by Enron employees Christian Yoder and Stephen Hall on December 6, 2000 concerning unjust and unreasonable trading and market manipulation.

Game 1: “Fat Boy” (Or “INCing”) The “Fat Boy” strategy described in the Yoder-Hall memo was designed to skirt the requirement that Scheduling Coordinators (SCs) were required to submit balanced supply and demand schedules to the CPX. Using this strategy, sellers would purposely submit false schedules to the CPX containing a load or demand schedule that was artificially inflated (“INCing” the load). Because there was actually a lower demand than identified on the schedule submitted to the CPX, the seller would be able to supply this excess energy to the CAISO’s real time or imbalance energy market. The game appears to have evolved in response to a game played by California IOUs where they under scheduled their demand in the CPX’s day-ahead market in an attempt to reduce prices in that market. 1

The California Parties include the Attorney General of the State of California, the California Public Utilities Commission, the Electricity Oversight Board, Southern California Edison, and Pacific Gas & Electric.

Market Manipulation

129

Several points are important. First, this game is hardly a new concept. The essence of “Fat Boy” is to buy low and sell high. Second, the CAISO was aware that market participants were overscheduling and underscheduling load. In fact, because overscheduling load in the CPX market meant that any unused electricity would be available in its own realtime imbalance market, the CAISO actually encouraged sellers to engage in overscheduling. Third, the unanticipated events that roiled California’s wholesale electricity market in late 2000 caused the Fat Boy strategy to grow to levels no one could have foreseen. Recall that when the CAISO was designed, it was anticipated that its real-time imbalance market would represent no more than 3 percent of the electricity market in California. However, due to the utility underscheduling in the CPX day-ahead market, upwards of 30 percent of the California market eventually went through the CAISO’s real-time imbalance market. These events caused the level and strategies in the Fat Boy game to change. Due to these events, it became very likely that the game would be very profitable. The representative or stakeholder boards that managed the CPX and CAISO were not then independent, as they are today. Consequently, these boards failed to act quickly and the CAISO’s real-time purchases soared. The FERC investigated these trading strategies in great detail to ascertain whether the market in California had been harmed by the Fat Boy strategy. In 2003, the FERC in its Order to Show Cause Concerning Gaming and/or Anomalous Market Behavior (Docket No. EL03-127-000)2 found that, although overscheduling load required filing false schedules with the CAISO, this sellers’ practice was initiated in response to the underscheduling by the buyers or IOUs. In effect, the FERC found both sellers and buyers guilty, but did not punish either side because their bidding strategies canceled each other out.

Game 2: “Death Star” (Or “Relieving Congestion and Counterflow Payments”) In this game, market participants make money by bidding to sell energy or biding to relieve congestion. This was accomplished by either selling against the primary direction of energy flow or by reducing demand. The rules and prices for congestion relief are extremely complex. Once the rules were understood, traders could submit bids that were within the literal wording of the rules and reap large profits for doing nothing. 2

Order to Show Cause Concerning Gaming and/or Anomalous Market Behavior, 103 FERC ¶61,345 (June 25, 2003).

130

The California Electricity Crisis: What, Why, and What’s Next

In essence, “Death Star” was designed to earn congestion revenues through differences in import/export schedules. This trading strategy involved scheduling an export out of the CAISO control area and a simultaneous import into the CAISO control area. In the CAISO’s transactions records, these trades would appear as two separate transactions. The CAISO alleged that, in reality, there was often a third transaction or schedule outside the CASIO control area that formed a closed loop between the import to and export from the CAISO control area. The party instituting this three-step strategy would earn congestion revenue by creating a counter-flow, although there was in actuality no specific physical beginning (source) or end (sink). No energy would actually flow anywhere because the schedules were circular and selfcanceling. Even though the CAISO had been pre-warned before the markets opened that this strategy likely would be profitable, the CAISO failed to act to close this opportunity until Enron made a trade where it tried to schedule an outlandish amount of electricity over the tiny Silver Peak line. The schedule created congestion that Enron “relieved,” for a fee, by scheduling a counterflow. To its credit, the CAISO finally fixed this flaw in its transmission tariff and fined Enron, which promised to never again submit another schedule like this one. The FERC staff issued a Show Cause Order in which traders needed to prove that they did not earn unreasonable income from transmission congestion relief. The proceeding was ongoing at the time we wrote this book. Some market participants have successfully explained their activities and have been dropped from this investigation.

Game 3: Load Shift This game is a form of arbitrage. Market participants would submit multiple bids in advance of a market’s close. Subsequently, based upon updated information (e.g., emergency warnings, weather conditions, etc.), the participants would alter their bids. All commodity markets work in this manner. Such arbitrage and hedging activities are essential for markets to be efficient. The California market tariffs, therefore, allowed this type of activity to take place. However, by 2000, there were several different price cap and other restrictions in place in the various CPX and CAISO markets. These price caps served to exacerbate the gaming behavior. For example, the CPX price cap was $2,500 per MWH. In contrast, the CAISO market cap was only $250 per MWH. Buyers could protect themselves by underscheduling in the CPX market, thereby shifting purchases to the CAISO’s market. Sellers

Market Manipulation

131

soon developed a response to the buyers’ practice, sometimes called “ricochet,” which we discuss in detail below. As noted above, these were completely unanticipated developments. Load shift continued until the spring of 2001, at which time the FERC imposed full regional western-state market participation requirements and bidding rules, which eventually led to regional price caps. These actions eliminated the incentives that precipitated this conduct.

Game 4: “Ricochet” (Or “Megawatt-Hour Laundering”) We mentioned this game above. The CAISO’s tariff provision capping the price of electricity sold in its markets combined with the CPX’s much higher market price cap to give rise to this game. Also important was the fact that certain entities, such as the municipal utilities and out-of-state generators, were not bound by the CAISO’s price cap because they were not market participants. Therefore, the price cap did not apply when non-exempt entities sold to exempt entities. Thus, the CAISO price cap could be circumvented by selling power to an exempt entity, which in turn, could sell that same power back to the CAISO unfettered by the CAISO’s price caps. Enron allegedly actively sought partners with which it could engage in these activities. It was not until late spring of 2001 that the FERC finally took action to eliminate this behavior.

Game 5: “Get Shorty” Two potential strategies were involved in this game. The first took advantage of the systematic differences in the day-ahead and hour-ahead market prices for ancillary services. Under “Get Shorty” trading, ancillary services would be sold in the day-ahead market, and then bought back in hour-ahead market at a price the seller hoped would be lower than the price in the hour-ahead market, thus earning the difference. The second strategy involved selling ancillary services in the day-ahead market from imports for which the resources are not actually available and then covering these sales by purchasing ancillary services in the hour-ahead market at what the buyer hoped would be a price lower than the price it sold the ancillary services for in the day-ahead market. This “game” is essentially a short-selling strategy that is a common practice among all commodity and stock traders. A trader agrees to buy or sell a product at a specified price in advance. As the date or time to execute the transaction approaches, the actual trading price’s value becomes more certain and the trader may change his/her position based upon the updated

132

The California Electricity Crisis: What, Why, and What’s Next

current price information. Enron did not invent this strategy. Rules in California against long-term contracts made the game very profitable. However, in order to offer ancillary services, the seller is required to 3 have the rights to the ancillary services it bids to sell. The FERC Trial Staff and California Parties are currently investigating trading practices to make certain that sellers actually owned the Ancillary Service that they bid. Some sellers have already been exonerated. The investigation continues into 2004.

Game 6: Wheel Out This game is all about information related to transmission line outages. Of course, there is nothing sinister in valuing information. The accusations, however, of wrongdoing are related to falsifying information. This strategy involved a Scheduling Coordinator (SC) submitting schedules to ship energy on out of service or de-rated (to zero) tie points. Doing so would require the CAISO to pay for a counter-flow schedule in real time because the CAISO’s software will automatically adjust schedules to achieve a zero flow across the de-rated line by accepting Adjustment Bids for counter-flows in the opposite direction in the same MW amount as the original schedule in order to get back to the zero rating. If the adjustment schedule is accepted, the SC would have been paid counter-flow revenues for relieving the fictitious congestion it caused. Thus, the SC can submit a schedule over an out-of-service line, and then submit an adjustment schedule, knowing that the CAISO will automatically cancel out both schedules and pay the SC counter-flow revenues for “relieving” the congestion. FERC Staff working with data from the California Parties have identified potentially illegal trades of this nature. Some SCs have responded and been exonerated. The investigation continues into 2004. Next, we discuss some questionable game strategies that were not discussed in the Enron memo.

Game 7: Selling Non-Firm Power as Firm Under this strategy, sellers would break the market rules by selling non-firm power imported from outside the CAISO control area into the CPX as firm energy. In this context, firm energy is energy that includes ancillary services. Thus, the seller is paid for ancillary services that it is not in fact providing.

3

See FERC Staff Motion to Dismiss in Docket No. EL03-145-000.

Market Manipulation

133

Game 8: Scheduling Energy to Collect Congestion Charges (“Cut Counter-Flows”) Prior to July 21, 2000, an SC could submit schedules in the day-ahead and/or hour-ahead congestion markets to provide counter-flows on congested paths. This would earn the SC congestion revenue, paid by the SCs who had submitted schedules in the congested direction. Under the CAISO rules, the SC could reduce its counter-flow schedule prior to real time but still receive the counter-flow revenues for schedules submitted in the day-ahead or hourahead congestion markets. This practice was identified and proscribed by the CAISO on July 21, 2000, which was before the Refund Period.4

Game 9: Anomalous Bidding Behavior This is not a game strategy. It is a prohibited activity. In its Order Requiring Demonstration That Certain Bids Did Not Constitute Anomalous Market Behavior; Docket No. IN03-10-009 (June 25, 2003), the FERC noted that the CAISO’s and CPX’s Market Monitoring and Information Protocols (MMIPs) prohibited “anomalous market behavior that departs significantly from normal behavior in a competitive market.”5 The FERC stated that this included “withholding of generation capacity under circumstances in which it would normally be offered at a competitive market.” The FERC also noted that the MMIP expressly prohibited “pricing and bidding patterns that are inconsistent with prevailing supply and demand conditions (e.g., prices and bids that appear consistently excessive for or otherwise inconsistent with such conditions”6). Thus, the FERC adopted a market wide screen recommended by FERC Staff that all bids in the CAISO and CPX markets above $250 per MW for the period May 1, 2000 to October 2, 2000 be considered excessive prima facie, and that those entities that submitted bids exceeding $250 per MW would be required, in a show cause investigation, to demonstrate why their bidding behavior and practices did not violate the MMIP and constituted legitimate business behavior. This analysis continues and will be discussed in detail below.

Game 10: Scheduling Services Arrangements The FERC alleged in its Order to Show Cause Concerning Gaming and/or Anomalous Market Behavior Through the Use of Partnerships, Alliances or 4

The Refund Period runs from October 2, 2000 through June 20, 2001. See MMIP Section 2.1.1.5. 6 Ibid. 5

134

The California Electricity Crisis: What, Why, and What’s Next

Other Arrangements, Docket No. EL03-180-00- et al. (June 25, 2003), that several entities (that they identify within the order) “worked in concert through partnerships, alliances, or other arrangements to engage in activities that constitute gaming or anomalous market behavior during the period January 1, 2000 to June 20, 2001.” Their primary focus was on Enron Power Marketing, Inc. and Enron Energy Services, Inc. The trading practices FERC identified were (1) False Import; (2) Congestion-Related Practices (Cutting Non-Firm, Circular Scheduling, Scheduling Counterflows on Out-of-Service Lines, and Load Shift); (3) Ancillary ServicesRelated Practices (Paper Trading and Double-Selling; and (4) Selling NonFirm Energy as Firm. The FERC concluded that although several of the entities involved are governmental entities, pursuant to its December 19, 2001 Order, governmental entities that sold energy in the CAISO and CPX short-term energy markets would be subject to refund liability. Focusing primarily on Enron, the FERC identified several partnerships.7 The FERC found that the arrangement normally started as consulting services that allowed entities to outsource certain tasks to Enron for a fee. These arrangements then evolved into a more comprehensive partnership when wholesale power marketing became involved and profits were split between the partners. Eventually Enron would “gain control of decision-making in a way that maximized profits for itself and its business partners.” Enron provided scheduling services for El Paso Electric, Glendale, CFE, Tosco, Washington Water Power, and Enron Energy Services. Each has been or is still being investigated. The Show Cause Order lists several other apparent partnerships, alliances, or other arrangements that may have engaged in market manipulation schemes. These include: (1) Sempra and Eugene Water and Electricity Board; (2) Coral and Glendale; (3) PSNM and Aquila, Constellation Power Source and El Paso Merchant, Enron, Idaho Power, Koch Energy Trading, MIECO, Morgan Stanley Capital Group, PECO, PacifiCorp, Powerex, Sempra, or TransAlta Energy Marketing. These arrangements are being or have been investigated. Settlements or penalty payments made without admitting guilt have been commonplace.

7

City of Glendale, City of Redding, Colorado River Association, Las Vegas Cogeneration, Modesto Irrigation District, Montana Power Company, Northern California Power Agency, Powerex Corporation, Public Service Company of New Mexico, and Valley Electric Association.

Market Manipulation

135

Game 11: Wash Trades In its March 2003 Report, FERC Staff defines wash trades as “a prearranged pair of trades of the same good between the same parties, involving no economic risk and no net change in beneficial ownership. The trades expose the parties to no monetary risk and serve no legitimate business purpose.”8 FERC Staff identified nine companies that comprised 57 percent of the wash trades. The balance of the potential wash trades (43 percent) was done by unidentified companies. Wash trades inflate revenue but not income. False signals are sent to financial analysts. Market shares and growth are overstated. This practice has also attracted Department of Justice and Securities and Exchange Commission review. Legal actions, penalties, shareholder and board of director actions have been taken.

Game 12: Affiliate Price Manipulation These are wash trades and other anomalous trades between affiliated companies. Enron has figured prominently in these investigations. These actions have been receiving ongoing review for obvious reasons related to Enron’s collapse.

CONCLUSION The investigations with respect to market manipulation are ongoing. To date, some charges have been dismissed and in others, penalties have been paid without admission of guilt. The process is not complete. Therefore, it is not possible to reach a final conclusion. That said, two things seem relatively certain. First, the root causes of California’s electricity crisis remain: market forces, climate, and design flaws. Market manipulation accusations, while troublesome, were mostly related to individual traders seeking to game the rules or to cheat. These actions work best when undetected. If the market moves, as it does with underscheduling, losers in the market will respond and regulators will identify and attempt to fix the problem. Second, energy markets need a strong and fast-acting market monitoring program. California recognized this need. However, the previous discussion and the fact that years later the investigation continues suggest 8

“Final Report on Price Manipulation in Western Markets: Fact-Finding Investigation of Potential Manipulation of Electric and Natural Gas Prices,” (prepared by the Staff of the Federal Energy Regulatory Commission), Docket No. PA02-2-000, (March 2003).

136

The California Electricity Crisis: What, Why, and What’s Next

that California’s monitoring function was not initially up to the task. We take up this subject in Chapter 14.

Chapter 13 GAMING AND CHEATING

Commodity markets are uncertain because supply and demand‚ weather‚ and their degree of interdependence with other markets and factors are constantly changing. In most commodity markets‚ players hedge their positions to reduce risk when facing uncertainty. The FERC and state regulators generally accept market prices as just and reasonable or fair when markets are competitive. In brief‚ this means that no entity or groups of entities can or do manipulate market prices. In such circumstances‚ all participants become price-takers. As described above‚ California’s electricity market required that market participants buy and sell their electricity in spot markets. The IOUs’ ability to hedge with long-term contracts or generation was severely restricted. This design was unique among restructured electricity markets. Most markets have a very limited spot market. This unique design proved to be a significant flaw in the California’s market design. We would expect that any other restructuring efforts around the world will take note of California’s meltdown and decline to impose such constraints on energy markets in the future. Complicated markets that require all-in participation raise the stakes for market manipulation‚ especially when the opportunity to hedge the game’s outcome through long-term bilateral contracts is denied. Two related concepts are important: gaming and cheating. First‚ we need to define what the “game” was in California. California’s game was a unique one in restructured electricity markets and it was based on the excess supply that existed in the marketplace when the restructuring plan was implemented. In essence‚ participants won or lost based on short-term savings in demand. Thus‚ the game was a complex one predicated on excess supply and multiple markets where trading between the CPX and the CAISO was encouraged.

138

The California Electricity Crisis: What‚ Why‚ and What’s Next

Within the context of the “game,” strategies, or “gaming,” are developed to improve a participant’s likelihood of success. In deep or highly liquid commodity markets, participants will find it difficult to engage in gaming strategies without being detected unless the market rules are very complex. California’s market design was extremely complex. Gaming was predicted, expected, and observed. Buyers and sellers attempted to game the system. Even the CAISO devised games it attempted to play. The key point here is that gaming is not inherently bad nor undesirable as long as the market participants stay within the rules and market overseers and monitors have the ability to change rules if certain gaming strategies lead to unfair or inefficient outcomes. Actions taken outside of the rules by market participants do not fall within the definition of gaming. Rather, it is cheating. Cheating is not strategic. Cheating violates the rules, threatens the market, and should be punished. It is often difficult in markets as complex as California to ascertain which activities are acceptable instances of strategic gaming and which activities cross the boundaries and become illegal cheating. The gut-wrenching California electricity crisis has created a political/regulatory morass where such judgments are still being sorted out. In order to try to bring some clarity to the debate, we next discuss three concepts: market forces, market power, and market gaming. Each concept is quite different, although the similarity in names sometimes confounds non-economists. Market Forces are supply and demand factors such as not building generation, droughts, air conditioning surges, etc. These forces are often uncertain. For example, the economic growth in California, particularly in high-tech areas during the late 1990s, was not anticipated. In California, as we discussed earlier, there were also significant increases in supply-side prices for inputs, such as a thirty-fold increase in natural gas prices in California and a twenty-fold increase in NOx air pollution compliance costs. Economists define Market Power (or monopoly power) as the ability of one seller or an illegal conspiracy of several sellers to withhold supply to force up prices. The consumer corollary is called monopsony power, and this occurs when buyers act to reduce demand to cause market prices to fall. We discussed underscheduling and overscheduling in Chapter 7. These trading activities represent monopoly and monopsony aspects of market power abuse. Market Gaming occurs when individual market participants engage in self-serving strategies (often legal and within the rules) that are generally contrary to the overall market. In contrast to those market power abusers who attempt to move the entire market, gamers employ strategies, in effect,

Gaming and Cheating

139

that bet against the overall market. Gaming is an individual activity. If all market participants employed the same strategy‚ there would be no opportunity to profit from beating the market. By their very nature‚ Market Forces cannot be constrained by laws or regulation. Attempts to regulate market forces are futile and always fail. This futility will not‚ however‚ deter some politicians from attempting to make such efforts or claiming success when markets shift favorably for constituents. Market Power is an entirely different matter. Potential and actual antitrust violations are harmful and must be prohibited. The standard here is simple: did sellers or buyers withhold quantities to force market prices up or down? Or‚ do they have enough market share or potential to do so if they chose to do so? Market Gaming is typical in complex commodity markets. The term has unfairly become pejorative. Market participants often engage in hedging activities. These strategic actions improve efficiency and help identify rules that need to be strengthened or modified. Market Gaming is a niche game‚ most successful when undetected. Once a successful strategy is detected‚ it will be imitated. Once imitated‚ the effectiveness of the strategy tends to evaporate. Strategic game playing is most often successfully executed in complex‚ multi-faceted markets like the restructured California electricity market. Gaming can morph from legal strategies and become illegal. Participants who cheat by playing outside the market’s rules and protocols must be punished severely enough so that any ill-gotten gains are forfeited. Criminal prosecution and imprisonment are appropriate in extreme situations. The distinction between “good” market-improving gaming and illegal gaming is the central issue for determining the extent and significance of market manipulation in California.

GAMING IN THE CONTEXT OF THE CALIFORNIA POWER MARKET Section 2.0 of the CAISO tariff‚ in incorporating the MMIP‚ states‚ “The ISO shall monitor the markets that it administers in order to identify and‚ where appropriate‚ institute corrective action to respond to the exercise of market power or other abuses of such markets...” The MMIP recognizes that there are likely to be design flaws and inefficiencies in the market design. MMIP Section 1.1 sets forth the objectives of the MMIP. It states:

140

The California Electricity Crisis: What‚ Why‚ and What’s Next

“This Protocol (MMIP) sets forth the workplan and‚ where applicable‚ the rules under which the ISO will monitor the ISO markets to identify abuses of market power‚ to ensure to the extent possible the efficient working of the ISO Markets immediately upon commencement of their operation‚ and to provide for their protection from abuses of market power in both the short term and the long term‚ and from other abuses that have the potential to undermine their effective functioning or overall efficiency...” The CAISO’s Market Surveillance Unit was initially established to monitor certain activities‚ including “anomalous market behavior” and “gaming.” MMIP Sections 2.1.4 (CAISO and CPX Design Flaws) and 2.1.5 (Market Structure Flaws) both evidence the fact that the market surveillance units were to use anomalous behavior and gaming to identify and correct design and market structural flaws. MMIP Section 2.3.1 allows the CAISO to take corrective action where the presence of gaming or exercises of market power would adversely affect the operation of the CAISO’s markets. The FERC Staff has implied that market participants had been put on notice that gaming was strictly prohibited. This is not correct. MMIP Section 2.3.3 made it clear that the CAISO’s Market Surveillance Unit‚ in cooperation with the CPX’s Compliance Unit‚ would use market monitoring to review the behavior of market participants. If anomalous behavior or gaming was found‚ the CAISO‚ through its Market Monitoring Unit‚ would: “...review the ‘gaming’ behavior and/or relationship between system conditions and market behavior and pricing in order to assess the potential for and impact of such gaming behavior‚ with a view to taking appropriate action‚ if necessary‚ wither with respect to structural changes such as Zone changes‚ or to changes to the ISO or PX Tariffs‚ Protocols or Activity Rules‚ or to proscribe specific behavior by Market Participants.” In the CAISO’s recent proposed amendment to the MMIP‚1 these basic provisions remain unchanged. Contrary to the implications of the FERC Staff‚ neither the CAISO tariff nor the MMIP expressly prohibited gaming. After first defining gaming behavior that takes “undue” or “unfair” advantage of the rules‚2 the CAISO tariff merely subjected gaming to 1

See 1/03/03 Version of Revisions for O&I review to ISO Market Monitoring & Information Protocol. 2 MMIP 2.1.4. Gaming is defined as taking unfair advantage of the rules and procedures set forth in the PX or CAISO tariffs, or of transmission constraints in periods where there is substantial congestion, to the detriment of and efficiency and consumers. Gaming, under this CAISO tariff provision, can also include taking undue advantage of other conditions

Gaming and Cheating

141

scrutiny. Even as the CAISO defined it‚ gaming behavior did not automatically lead to the imposition of remedies. Instead‚ the CAISO tariff authorized its Market Surveillance Unit (MSU) to review gaming behavior in order to assess its potential effect. Such assessments could result in recommendations by the MSU to make structural changes‚ to make tariff changes‚ or to proscribe specific behavior.3 This is consistent with the notion that the market system being put into place was known not to be perfect and that it would be “fixed” as problems were identified. The CAISO and the CPX intended to rely on market monitors to identify behavior that could negatively affect the markets‚ and then could take action to remedy the problem. The CAISO tariff‚ in effect‚ recognized that gaming is not inherently bad. Indeed‚ “arbitrage‚” which is defined as buying low and selling high across markets‚ is good and promotes efficiency in markets. Commodity markets are monitored in order to revise rules that result in inefficient or unreasonable gaming. All commodity markets are gamed in the sense that buyers and sellers adopt‚ refine‚ and revise their bidding strategies based upon the actual or anticipated behavior of others. Buyers and sellers often hedge their bets. Their actions improve efficiency and help to identify any fundamental structural changes. The CAISO FERC-approved tariff underscores the valuable and necessary function these natural market activities play in a commodities market–and how the CAISO itself recognized that gaming could constitute legitimate aggressive competition. Market “manipulation‚” or cheating‚ means conduct that is prohibited by market rules or law in which a market participant engages in an activity with the intention of influencing prices and that results in creating an undue disadvantage for other market participants. Our previous comments about market gaming should not be applied to market manipulation or cheating. There is no doubt that natural gas and electricity prices jumped dramatically over a fourteen-month period from May 2000 through June 2001. The FERC is reviewing evidence indicative or counter-indicative of market manipulation. This evidence relates to some form of unlawful withholding‚ price fixing‚ or gaming outside the regulatory rules in place at the time. The challenge is to untangle the various factors that cause prices to increase in order to determine the extent of price increases due to market manipulation or illegal activities. Games are often confused with market

3

that may affect the availability of transmission and generation capacity‚ such as loop flow‚ facility outages‚ hydropower output levels and seasonal limits on out-of-state energy imports‚ or actions or behaviors that may otherwise render the system and the CAISO markets vulnerable to price manipulation to the detriment of their efficiency. MMIP 2.3.3 (“Response to Gaming Behavior”).

142

The California Electricity Crisis: What‚ Why‚ and What’s Next

power. Market forces that cause prices to increase and/or vary should not be ignored.

REGULATION AND GAMING Structural design flaws also existed in the California market. This leads to some important considerations. First‚ suppliers would not sell goods at their marginal production cost when there are shortages because marginal opportunity costs take on greater importance. If sellers have marginal costs less than or equal to the market price‚ they would bid their goods at the expected market price‚ including opportunity cost‚ not their marginal production cost. Next‚ spot markets are a somewhat special case because the period that a particular price remains in effect is relatively short (e.g.‚ an hour or a day). Spot market bidders sometimes will rationally bid zero to be guaranteed that they will be “in the money.” This is a reasonable approach‚ not because marginal production costs are really zero‚ but because the uncertain market clearing prices are likely to be significantly greater than the seller’s marginal production costs and these sellers want to supply the market in that hour. Finally‚ there is also rationale for the opposite bidding strategy of super high bids. It is erroneous to assume that high bids and low marginal costs automatically spell monopoly power or collusion. Consider a person that purchases a ticket to an event such as a basketball game for $10. Now suppose the event becomes sold out and it becomes reported that tickets are now selling for $200. Now assume that this person cannot attend the event. Would he or she seek a $10 refund‚ the marginal purchase cost‚ or seek to sell the ticket for a price in the reported neighborhood of $200 to reflect their marginal opportunity cost? In competitive markets with no restrictions on secondary markets for reselling‚ the person would seek a market price of about $200‚ not the $10 he or she paid for the ticket. Now‚ suppose the ticket seller thinks prices could go higher. There is nothing sinister about a seller putting out a $250 offer to sell this same $10 ticket if the seller thinks that as show time approaches‚ the ticket’s value traded in real time could be even higher than the $200 advance market’s single price. Under such conditions‚ the advance market price could clear at or above $250. Suppose it did not clear at $250‚ the seller could still arrange a real-time sale and recoup a profit over the $10. Such profit could be less than‚ equal to‚ or greater than the $200 bid/ask in the advance-ticket secondary market. There are no competitive-market bidding rules here. Guts‚ risk-taking‚ and experience drive behavior.

Gaming and Cheating

143

It is reasonable to expect that participants in competitive markets will explore ways in which to best use the rules that are established in the market. The competitive wholesale power market established in California was no different in this expectation. This was why the MMIP was established to monitor the market’s function and identify “gaming” and “anomalous market behavior.” This was also true for the CPX’s Compliance Unit‚ which issued periodic monthly and quarterly reports on CPX market activity‚ and its Market Monitoring Committee‚ an outside group of academic experts who independently assessed the issues and trends affecting the market. The CAISO formed its Department of Market Analysis (DMA) to perform functions parallel to the CPX’s Compliance Unit. The CAISO also created the Market Surveillance Committee to provide independent reporting on market issues and trends. These monitoring functions indicate that the CAISO and CPX expected and anticipated that market participants would attempt to game the system and that there would be situations that might demonstrate that anomalous behavior or market power was being exercised. Gaming is a natural phenomenon in competitive markets and is useful to identify weaknesses and structural flaws in markets‚ especially markets that are in their infancy. Before the first MWH was sold‚ some who were involved in the process said that there would be problems with the market. However‚ rather than delay the implementation of the market in California by trying to get the system “perfect‚” legislation was enacted with the understanding that there would be a shake-out period where flaws and problems were identified and corrected. This was one reason why the CPX and CAISO formed market monitoring groups. In fact‚ the CPX and CAISO market monitoring groups identified behavior that they considered to be troubling as early as March 1999‚ when the CAISO Surveillance Committee and CPX Market Monitoring Committee reported on underscheduling of demand in the CPX day-ahead market. These concerns‚ along with others related to higher than expected demand for ancillary services‚ underscheduling supply in the CPX day-ahead market‚ the problematic nature of the CAISO and CPX markets’ sequential structure‚ the lack of price signals to consumers‚ and the prohibition from contracting outside of the CPX day-ahead market‚ were routinely reported to the CAISO and CPX management and boards. Generally‚ no legislative or regulatory action was taken in 1999 and 2000. Additionally‚ the stakeholder composition of the CAISO board often prevented effective corrective action from being initiated. On occasion‚ the market monitoring worked as intended. For example‚ in the now infamous Silver Peak incident we discussed earlier where Enron “gamed” the system by vastly overscheduling load on a single line in order

144

The California Electricity Crisis: What‚ Why‚ and What’s Next

to collect congestion relief payments‚ the CAISO quickly identified the “game” and fined Enron for its behavior‚ also extracting Enron’s promise to never again attempt such a strategy in the future. This is exactly how the MMIP was supposed to function. Gaming or anomalous behavior would be identified and steps would be taken to correct the situation. Often‚ flaws in the market are discoverable only by observing how the market and the market participants are performing. In these situations‚ gaming by market participants is expected and necessary to identify areas where the market rules can be fixed and/or strengthened. Consequently‚ the legislators and regulators who approved the CAISO and CPX tariffs contemplated gaming and arbitrage‚ because that is expected behavior in all competitive markets. However‚ legislators and regulators also recognized that market surveillance was necessary to identify these behaviors and recommend to the respective boards‚ legislators‚ and regulatory agencies tariff modifications that would address those issues. The CAISO and CPX were controlled by a stakeholder board. Most of the concerns identified by the market monitoring groups at the CAISO and the CPX were ignored or tabled until it was too late to avert the crisis in California. FERC‚ which had no independent market monitoring system within California‚ was largely kept in the dark. It is important to distinguish between behavior that was within the rules established by the tariff and behavior that was illegal or outside the rules. The market monitoring groups were tasked with and did uncover behavior that was within the rules‚ but which required some tariff modification to correct. This was all anticipated by those who approved the tariff.

CONCLUSION Gaming behavior was not limited to the IOUs‚ generators‚ and other market participants. The CAISO itself attempted to game the system. The CAISO’s CEO Terry Winter has admitted that the CAISO falsely created the illusion of congestion on its lines in order to prevent electricity from flowing from California. In fact‚ in Congressional hearings‚ Congressman Ose commended Mr. Winter for taking this “gaming” action to protect California ratepayers. While the CAISO’s motivation was certainly different‚ it is hard to distinguish‚ from a “gaming” perspective‚ Enron’s Silver Peak congestion “game” from the CAISO’s phony-congestion-to-prevent-power-outflow “game.” Both games attempted to take advantage of the rules. Enron caused the rules to be changed and had to pay a fine. The CAISO performed well to reduce costs under exigent circumstances that made the existing rules rigid.

Gaming and Cheating

145

The CAISO also attempted to “game” the system after the FERC imposed a western states price cap that could be reset during electricity emergencies. It appears the CAISO falsely declared an emergency when there was none. The CAISO’s purpose in this “game” was to reset the price cap in California to a lower number. The FERC prevented the CAISO from running this “game” and prohibited such gaming behavior in the future. Gaming behavior was anticipated‚ expected‚ and required in order to identify weaknesses in the market design. Market monitoring groups were established to uncover these games played within the market rules and market design flaws. These monitoring groups were expected to report their finding to the boards so that corrective action could be taken by modifying the tariffs. Everything performed as anticipated‚ except that regulators and stakeholder boards‚ for whatever reason‚ initially failed to take swift corrective action. After the Enron fiasco‚ years spent reviewing refund liability‚ and individual company show-cause investigations‚ California’s market “experiment” continues to operate. Some assert that they will not participate in markets where the players must “strike out the umpire” in order to be successful. Market monitoring and corrective regulation are necessary because it is unreasonable to think that legislators and regulators will get all the rules straight from the outset. Failing to correct flaws expeditiously seems even more unreasonable. This is certainly true when “putting things right” means that markets need to be recreated through ex post simulation years after they closed. As California has proven‚ such action is extraordinarily difficult. In Chapters 15 and 16‚ we investigate California refund process in some detail. But first‚ in Chapter 14‚ we review the role of market monitors and the initial regulatory and political reaction to the initial calls for reform.

This page intentionally left blank

Chapter 14 MARKET MONITORING AND INITIAL REGULATORY RESPONSE

Deregulating wholesale electricity markets and the need for market monitoring are commingled. This relationship has essential service and franchise utility roots. California’s experience has given additional meaning to this interdependence between liberalization and scrutiny to assure the rules are followed and to change the markets’ rules if there are unreasonable outcomes. California’s market design was codified in FERC-approved tariffs, which address gaming and strategic behavior. For example, as we discussed in Chapter 13, the CPX and CAISO markets were required to have interval and interdependent market monitors that were required to analyze market clearing prices searching for market power and unreasonable manipulation, as we discussed in the previous chapter. Also, the FERC Tariffs defined trading behavior that was not permitted. Section 2.0 of the CAISO Tariff incorporates Market Monitoring and Information Protocol (MMIP) and states, “The ISO shall monitor the markets that it administers in order to identify and, where appropriate, institute corrective action to respond to the exercise of market power or other abuses of such markets...” Also, the CAISO, CPX, and FERC fully anticipated that when problems developed and the market matured, the tariffs, (i.e., trading rules), would be modified and further restructuring was fully contemplated. This chapter reviews how market monitoring worked during the California Electricity Crisis. These monitoring groups warned the CPUC and the FERC of potential problems caused by underscheduling, increased demand, and the mandated use of short-term markets. As we noted previously, market-monitoring groups within the CAISO and CPX began to

148

The California Electricity Crisis: What, Why, and What’s Next

identify problems in the electricity market structure as early as the fall of 1998.

HISTORY OF CAISO AND PX MARKET MONITORING Both the CAISO and CPX established special market-monitoring units in addition to committees of outside experts to provide an independent assessment of issues that would affect the market. The CPX hired a relatively small staff, called the Compliance Unit, to issue periodic reports (both monthly and quarterly) on CPX market activity. It also assembled a group of outside academic experts as its Market Monitoring Committee, intended to provide an independent assessment of issues and trends that would affect the market. In addition to its quarterly meetings, the Market Monitoring Committee was required to submit an annual report concerning market issues to FERC. The CAISO designed its Department of Market Analysis to perform functions parallel to those of the Compliance Unit and created the Market Surveillance Committee to provide independent reporting on market issues and trends. As early as August 1998 both the CAISO and CPX monitoring groups identified anomalous market behavior and structural design flaws. Table 14-1 outlines several issues the CAISO and CPX market-monitoring groups identified in their various reports. For example, between March and October of 1999, the CAISO’s surveillance committee issued reports outlining what it believed to be key market concerns. The prominent concerns were: (1) underscheduling in the CPX day-ahead market; (2) retail consumers could not modify their demand for energy in response to wholesale prices; (3) the market’s overreliance on spot market energy purchases. Similarly, in March 1999, the CPX concluded that during periods of very high demand, a small number of power generators had the ability to set the price for wholesale electricity. The monitoring groups raised red flags by identifying and reporting these concerns.

Market Monitoring and Initial Regulatory Response

149

The market-monitoring groups were responsible for reporting market concerns to senior management or to their respective boards and then to

150

The California Electricity Crisis: What, Why, and What’s Next

external agencies as required. The CPX and the CAISO monitoring groups, between August 1998 and March 2000, contacted FERC and the CPUC on several occasions with respect to the flaws they perceived in the market’s structure. Sometimes, FERC and/or the CPUC responded by making changes. However, for the most part, FERC and the CPUC’s procedure rules for tariff revisions slowed the process significantly.

REGULATORY RESPONSES PRE-CRISIS Although FERC and the CPUC made changes to the market requirements, they were not successful in fully addressing and correcting concerns raised by market-monitoring groups. A case in point was the excessive reliance on the spot market. In July 1999, the CPUC issued a decision allowing two of the state’s IOUs to buy energy through forward contracts. The CPUC required that the utilities make these purchases only through the CPX, and limited the amount of energy that could be purchased under forward contracts. In making this decision, the CPUC stated that it was limiting the amount and duration of the power that could be secured using forward contracts because it was concerned that the IOUs would use forward contracts to engage in speculative trading in the CPX market. The CPUC’s July 1999 decision did not go far enough in providing the IOUs with the price protection they sought, because it had to act again in March 2000 to increase the amount of electricity the IOUs could purchase in the forward market. Nonetheless, with the likely possibility of reasonableness reviews, the IOUs did not avail themselves fully of their ability to enter into forward contracts. The CPUC acted again in August 2000 to grant IOUs the flexibility to enter into bilateral contracts. By that time, severe wholesale price spikes were already affecting the marketplace and utility credit worthiness became a new problem. Similarly, some of FERC’s actions failed to fully address and correct problems identified by the CPX and CAISO monitoring groups. For example, in November 2000, FERC acknowledged that a change it had approved in May 1999 to the CAISO’s procedures for procuring replacement reserves proved ineffective for reducing the amount of energy underscheduled in the CPX markets. These underscheduling strategies exacerbated the state’s electricity crisis in 2000.

Market Monitoring and Initial Regulatory Response

151

THE FERC’S INITIAL RESPONSE TO THE ONSET OF PRICE SPIKES On August 23, 2000, after California’s wholesale electricity markets experienced huge price increases, FERC commissioned an investigation by its staff into the conditions affecting California’s power markets. The results of this staff investigation prompted FERC to issue a preliminary order in November 2000 and a final order in December 2000 that outlined the changes it thought were necessary to restore order to the state’s wholesale electricity markets. Despite evidence suggesting that sellers during the summer of 2000 had the potential to exercise market power-defined by FERC as the ability of an entity to influence market outcomes for a sustained period—that may have contributed to higher wholesale electricity prices, FERC concluded that the evidence analyzed during its investigation was inconclusive in determining whether individual sellers exercised actual market power. FERC stated that further study of high bids made by individual firms or information concerning periods when generators were not running would be needed to substantiate any charges of market power abuse.

Detecting Unreasonable Behavior In the FERC’s preliminary order, issued in November 2000, the Commission concluded that “While the record did not support findings of specific exercises of market power in these spot markets, and while we were not able to reach definite conclusions about the actions of individual sellers, there was clear evidence that the California market structure and rules provide the opportunity for sellers to exercise market power when supply is tight and can result in unjust and unreasonable rates under the [Federal Power Act].” In response to this statement, various generators contacted the FERC to challenge these findings or demand clarification. In its final order, the FERC concluded that California’s electricity rates during the summer of 2000 were unjust and unreasonable under the Federal Power Act. However, responding to the comments of one generator (Dynegy) FERC also found that when analyzing rates for reasonableness, it could not look at an isolated time period but must look instead at a representative time period. “In response to Dynegy, we agree that in analyzing the reasonableness of rates in a particular market we cannot look at prices based on an isolated

152

The California Electricity Crisis: What, Why, and What’s Next

time period, but rather must look at a representative time period. We further agree that we need to distinguish scarcity rents from exercises of market power; however, we disagree that, absent exercise of market power, prices are necessarily just and reasonable. Our analysis must be, as discussed above, based on a determination of whether the rate falls within a zone of reasonableness.” Even with its finding that the rates were not just and reasonable, because of the complexity of the task, FERC informed the Western Governors Association in late December 2000 that it would not trace the dollars related to the transactions that occurred in the summer of 2000. At the meeting, the former chair of FERC stated, “Because any attempt to trace all of these dollars would be a time-consuming, if not impossible, exercise, the [FERC] has focused its efforts in western markets on operating procedures and fixing structural flaws in the underlying market design.” Moreover, as FERC outlined in its December 2000 order, it decided to limit the time it would use to analyze future transactions for assessing any potential refunds. “We clarify that, unless the [FERC] issues some form of notification to a seller that its transaction is still under review, refund potential on a particular transaction will close 60 days after the initial report1 is filed with the [FERC]. The institution of a 60-day period for the review of the transactions will provide sellers with the certainty they request and allows a reasonable period for analysis by staff.” In effect, the FERC punted as the crisis expanded in California, partially because FERC concluded that California’s markets were still developing. Even though FERC stated previously that it believed that summer 2000 wholesale electricity prices were unjust and unreasonable, FERC informed the Western Governors Association in late December 2000 that it would not pursue individual power generator refunds to those who purchased highpriced wholesale power because tracing the transactions would be a timeconsuming, if not impossible, task. The FERC reversed their position as the crisis grew and evidence of wrongdoing, such as the Enron Memo, became known. In December 2000, the FERC concluded that “the deregulatory approach adopted by California not only failed to address many of the existing problems that were plaguing the state, but in many ways the approach exacerbated and magnified those problems.” FERC concluded that, “The electric market structure and market rules for wholesale sales of 1

The initial report refers to FERC’s requirement that generators that are paid more than $150 per megawatt hour in the CPX or CAISO markets, report to FERC information on their costs of production.

Market Monitoring and Initial Regulatory Response

153

electric energy in California are seriously flawed.” In an attempt to rectify these problems, FERC, among other things, mandated certain actions in its December 15, 2000 Order.

Market Power In its December 15, 2000 Order, the FERC eliminated the requirement that the IOUs buy and sell all electricity through the CPX short-term markets. Despite evidence suggesting that sellers had the potential to exercise market power, FERC stated that further study would be needed to substantiate charges of market abuse, but did not follow through with such a study. The December 15 Order also forced the CAISO to restructure its governing board to remove stakeholders and replace them with members who were independent of market participants. Additionally, the December 15 Order allowed the CAISO to impose penalties when 95 percent of an entity’s demand was not scheduled a day ahead. This put an end to the IOUs’ underscheduling game. While these changes represented a step in the right direction, they did not fully address the flaws in the market structure. For example, the CAISO surveillance committee found that an abuse of market power occurred between October 1, 1999 through June 30, 2000: “The [surveillance committee] estimated a significant degree of market power being exercised in California markets for the period October 1, 1999, to June 30, 2000. . . . For the last month of the sample, June 2000, they estimated that prices were 64.6 percent higher than they would have been under competitive conditions. The highest previous monthly market power index was in June 1998, when prices were estimated to be 39.9 percent higher than they would have been under competitive conditions. These [findings] certainly suggest that market power was exercised in June by the standard of short run marginal costs.” In November 2000, the FERC acknowledged that the CAISO surveillance committee’s findings suggested that market power was exercised in June 2000. However, the FERC chose not to further investigate the potential market power abuse or to impose sanctions. In explaining its decision, the FERC noted the difficulty in separating higher prices due to the payment of scarcity premiums from the exercise of market power, as well as the difficulty in proving market abuse by individual firms. FERC argued that a power plant owner could exercise market power either by submitting bids that significantly exceeded its opportunity cost in order to raise the market-clearing price or by physically withholding power from the market in order to decrease the available supply, and that determining either type of market power abuse was problematic. The FERC also stated that generators

154

The California Electricity Crisis: What, Why, and What’s Next

had many different markets to choose from in California. Assessing true opportunity cost (the difference between the revenue gained and the revenue forfeited by rejecting alternative courses of action) would be difficult, and during periods of high demand, it would be difficult to differentiate physical withholding from real unit outages. Thus, what appeared to be an exercise of market power in June 2000 could have been a power plant owner responding to price signals or operational considerations. The FERC also observed that market problems may have been exacerbated by market design and implementation flaws. For this reason, the FERC stated that the best solution was to change the market rules. “Significant market power abuses that violate market rules need to be dealt with directly, but market power in a newly developing market may be magnified by flaws in market rules. The best approach in these cases may be to change the rules in order to mitigate the impact of market power exercise. Mitigation in the form of rule changes may be appropriate even in the absence of findings of market power exercise by specific sellers or buyers, if there are clear incentives for its exercise, and there are potentially large impacts that cannot be adequately separated from the effects of scarcity.”

MWH Laundering FERC also addressed MWH laundering and price cap avoidance and concluded in its December 2000 order that the strategy becomes a problem if done to drive up prices. Because there were no specific administrative rules requiring sellers to participate in the CPX market, such strategies were not found to be improper in December of 2000. According to FERC: “In one sense, this is not [improper] since there are no administrative rules on the amount of capacity (electricity) that must be provided to meet [demand] as there are in the eastern ISOs. [Buyers] are required to bid into the PX, but there is no capacity (seller) penalty imposed if corresponding supply does not bid into the PX.” According to the FERC order, “These exporting practices are permitted under the rules and are not necessarily a market power problem. It may simply be the normal working of a market where sellers are maximizing profits in a competitive market, where sellers or buyers see an opportunity at one time, take an option, and exercise it at a later date.”

Market Monitoring and Initial Regulatory Response

155

FERC went on to opine that, “It (MWH laundering) becomes a problem if it is associated with a pattern of withholding resources from the market in order to drive up prices. For example, if a large seller outside California were able to influence the price of power in the West by acquiring power from California, withholding power from the market at a critical time, and [then] selling the power back to California. As such, it is part of the overall issue of market power and scarcity in the West. . .”

Relax Spot-Market Purchase Requirements FERC’s most significant action in its December 2000 Order was to eliminate the requirement that megawatt hours needed to be sold in the spot markets. The original terms of deregulation required that IOUs make all of their energy purchases and sales in the CPX’s short-term markets. FERC eliminated this restriction, enabling the IOUs both to enter into new bilateral or long-term contracts to meet demand and to use the 25,000 megawatts (MWH) of generation they had not divested to serve their own customers. Under FERC’s order, the CPUC was responsible for approving the prices and terms of these new contracts. FERC adopted, for one year, a benchmark price of $74 per megawatt hour (MWH) that it would use to monitor whether rates for five-year wholesale supply contracts for electricity (forward contracts) were just and reasonable. However, FERC acknowledged that the benchmark was only intended to provide guidance to market participants and the CPUC.

Penalties on Buyers that Underscheduled FERC’s December Order also required IOUs to schedule 95 percent of their demand through either forward contracts or the day-ahead market. If the IOUs purchased more than 5 percent of the energy in the CAISO’s real-time energy market for any hour, FERC authorized the CAISO to impose a monetary penalty equal to two times the CAISO’s real-time cost of energy, not to exceed $100 per MWH, with the penalty revenues disbursed among those market participants that accurately scheduled 95 percent or more of their load.

156

The California Electricity Crisis: What, Why, and What’s Next

CONCLUSIONS The following story fits the observed facts. Others might tell a different story. Of course, what transpired in the minds of market monitors and regulators must be written by them. The story we think applies is that initially legislators and regulators were quite pleased with California’s relatively fast and brilliantly successful start. Recall the historic costs of operation were $50 per MWH or more, and that market-clearing prices during 1998 and 1999 averaged about half of their cost of service benchmark. In fact, market prices were low enough that the state’s smallest IOU, SDG&E, fully recovered its authorized stranded costs and was freed from the retail rate freeze required under AB 1890. The other two IOUs made statements in early 2000 that they would likely also beat the four-year recovery deadline. Market monitors were doubtlessly pleased with these happenings. Their job was to detect flaws, rules that needed to be changed, and anomalous trading activity. Their analyses and concerns were passed up the chain of review and made public. Both the CAISO and CPX Boards of Directors had what were termed Stakeholder Boards. These were mostly employees of special interest groups that helped organize and structure the new markets. For the most part. Board members, regulators, and legislators who were feeling the bright sun of success were not eager to listen to complex analyses that markets needed fixing. Doubtless, some concluded, “If it ain’t broke, don’t fix it.” Prices surged in the summer of 2000. However, the notion of success was still difficult to suppress. The CPUC and FERC were slow to act. In the fall of 2000, FERC was still expressing profound faith in the market and reaffirming the obvious fact the market forces work in both directions, pushing prices down and up. After first denying the need to fix the markets, regulators finally began to tweak them. As the crisis expanded and the normally slack fall months failed to stem the upward march of prices, both the FERC and the CPUC began to act. Seemingly, when their initial fixes, such as providing for forward contracts or discouraging underscheduling, did not work, regulators took to heart the market monitors constant and prolonged call to act. These actions beginning in early 2001 will be discussed in the following chapters. Here, we conclude this chapter by recapping the lessons of the monitoring and regulatory story. Two interrelated questions are important. The first explores the reasons supporting restructuring vertically integrated utility monopolies and replacing them with competitive wholesale-generation markets. The corollary explains the arguments for maintaining the status quo.

Market Monitoring and Initial Regulatory Response

157

There are four reasons to change from a vertically integrated monopoly. Each reflects economic and financial concerns. Prices are high, capital is scarce, and competition leads to more choices and improved economic efficiency. There are, of course, counterarguments. These include the time and effort required, the political difficulty of restructuring, and, of course, getting it right. Additionally, there is the sticky fact that markets work in both directions, causing political fallout and hand wringing when prices go up. Also, it is exorbitantly costly, both politically and economically, to try to put the genie back into the bottle through re-regulation. The last two reasons focus on what the last two chapters have been about. First, market clearing prices will increase when market forces push in that direction, and this will eventually happen. Second, monitoring, reforming, restructuring, and new regulation is difficult politically and economically. California and others took the first option and restructured. The California experience is still a work in progress. Some important lessons have been learned. The first lesson is to be wary of restructuring when there is excess supply because the excess supply will make restructuring appear to be easier than it really is. The excess supply will initially guarantee lower competitive prices relative to the incumbent IOU’s COS prices. Excess supply initially creates a win-win situation where it is easy to postpone making hard choices. Finally, excess supply will mask market flaws, hiding the need for regulatory reforms. The key lesson here is not to assume that market monitors will identify weaknesses in the market structure or that disclosures alone will be sufficient to fix market design flaws that are exposed when excess supply is replaces with shortages. The second lesson to be learned is that the system must be designed realizing that changes and modifications will be necessary. An independent procedure to quickly fix rules outside the influence and involvement of stakeholders is crucial. In other words, regulatory responsibility must be established. The third lesson is that market monitors must have sufficient authority to immediately penalize rule violators and to modify the rules in exigent circumstances. Market monitors must be more than data-collectors and analysts. Market monitors must be given clear policing authority and the ability to clearly define and establish the rules of the game. California was deep into designing a highly complex system that proved difficult to mend. Problems were identified that could, surprisingly, not be fixed. The fallout continues. Others need to learn the lessons of California. The three most important lessons are that: market prices will increase at times, market designs are not sacrosanct, and there must be procedures to monitor and change as problems, cheating, and anomalous behavior occur.

This page intentionally left blank

Chapter 15 REFUNDS AND MITIGATION

The FERC found that serious flaws existed in the California wholesale electricity market in November of 2000‚1 and further concluded that market clearing prices were unjust and unreasonable. As we noted earlier‚ in December 2000‚ the FERC took specific steps to mitigate prices. These included: (1) penalties for buyer underscheduling; (2) relaxing the IOUs’ requirements to purchase energy in the spot markets; (3) prospective refunds from generators that could not justify production cost basis prices that exceeded $150 per MWH; and‚ (4) calls for an end to stakeholder boards for the CPX and CAISO Boards of Directors. At that time‚ the general understanding was that the worst was over and that refunds would not be paid. However‚ the week of December 9 through December 16‚ 2000‚ had the worst sustained fly-up in market-clearing prices yet seen. Two IOUs (PG&E and SCE) and the CPX were on the brink of bankruptcy by January 2001. The state was forced to become the purchaser of MWHs in order to restore creditworthiness to the market. California’s problems‚ however‚ took a back seat to the presidential election controversy in Florida and the subsequent legal machinations that grabbed the nation’s attention. As the country settled down and California’s crisis continued‚ the FERC determined that it could establish a marginal cost benchmark for just and reasonable prices. This was a significant turning point because‚ in effect‚ the FERC viewed marginal cost as the marginal running cost of the least-efficient generator dispatched. This concept omits marginal opportunity cost‚ except as returns above the infra marginal generator’s marginal running costs.

1

Order Proposing Remedies for California Wholesale Markets‚ 93 FERC P61,121 at 61‚349 (2000) (November 1 Order).

160

The California Electricity Crisis: What‚ Why‚ and What’s Next

Significant regulatory policies were based on using proxy estimates of short-run marginal cost. First‚ the FERC established price caps using this benchmark. These are discussed next. Second‚ the FERC adopted these same proxy short-run marginal costs to calculate and order refunds for the refund period (October 2‚ 2000 to June 20‚ 2001). The basis for this refund and its time period are also discussed later in this chapter.

PRICE CAPS AND REGULATION In order to fully understand the FERC’s logic‚ it is important to analyze and understand several FERC orders in the Refund Proceeding. On August 23‚ 2000‚ the FERC issued an order commencing an investigation into California’s bulk power markets. 2 That investigation is still ongoing. This order is very important because it established the date the FERC ultimately found to be the legal starting point for the refund period. The FERC subsequently issued a series of orders concerning hard and soft price caps designed to control spot electricity prices‚ first in the California and subsequently all the western states. On December 15‚ 2000‚3 the FERC instituted a $150/MWH breakpoint methodology‚ which became effective January 1‚ 2001. Under the breakpoint methodology‚ sellers bidding at or below $150/MWH in the CAISO or CPX markets would receive the MCP‚ but not more than the $150/MWH breakpoint. Sellers bidding above the $150/MWH breakpoint would receive their bid prices if the energy was needed‚ but would be subject to weekly monitoring and reporting requirements to ensure that their bid prices were just and reasonable. This policy is called a soft cap because‚ while bid prices can exceed the $150/MWH soft cap‚ those prices would not set the single or MCP for other spot market sales. On April 26‚ 2001‚4 the FERC adopted a spot electricity marketmonitoring and price-mitigation plan to replace the $150/MWH breakpoint plan it had adopted earlier. The April 26‚ 2001 Order’s soft price cap applied whenever a Stage 1 emergency was declared (i.e.‚ when reserves 2

Order Initiating Hearing Proceedings to Investigate Justness and Reasonableness of Rates of Public Utility Sellers in California ISO and PX Markets and to Investigate ISO and PX Tariffs‚ Contracts‚ Institutional Structures and Bylaws; And Providing Further Guidance to California Entities; 92 FERC ¶61,172 (August 23‚ 2000). 3 Order Directing Remedies for California Wholesale Electric Markets‚ 93 FERC ¶61,294 (December 15‚ 2000). 4 Order Establishing Prospective Mitigation and Monitoring Plan for the California Wholesale Electric Markets and Establishing an Investigation of Public Utility Rates in Wholesale Western Energy Markets; 95 FERC ¶61,115 (April 26‚ 2001).

Refunds and Mitigation

161

were 7.5 percent5 or less). Prices were to be tied to marginal running cost. The FERC required each natural gas-fired generator in California to file its specific heat rate and emission information‚ which would be combined with generic estimates of operating expenses and market prices for natural gas. This information‚ in effect‚ would yield short-run marginal cost estimates that could serve as proxy prices. Generators could submit bids higher than their proxy price‚ but those bids could not and would not set the single spot MCP‚ and would be subject to FERC review and potential refund liability. Bids from outside California could not set the MCP because generators outside California did not provide the necessary proxy price data. This Order establishing the FERC price cap methodology ultimately provided the basis for the method FERC adopted for refunds. The April 26‚ 2001 Order also announced that the FERC was instituting a west-wide investigation under Section 206 of the Federal Power Act (FPA) into the rates‚ terms‚ and conditions of public utility sales for resale of electric energy in the entire western region or Western States Coordinating Council (WSCC). The Order also stated that any sales made in other realtime spot markets in the WSCC “would also be subject to price mitigation.” The FERC next issued an order on June 19‚ 2001‚6 which modified and clarified the April 26‚ 2001 Order. This Order effectively ended the electricity crisis in California by establishing western regional price caps for spot electricity market sales. This order ended MWH laundering and most interstate gaming strategies. This cap applies to all hours‚ not just reserve deficiency emergencies. The FERC’s new refund price cap is equal to 85 percent of the spot market clearing price during the last Stage 1 emergency period. This price cap would be the maximum permitted spot price‚ absent justification. This price was‚ therefore‚ set by the last increment of load during the prior Stage 1 reserve deficiency. Subsequent Stage 1 emergencies would cause the maximum price cap to be reset for subsequent non-reserve deficiency hourly prices at 85 percent of this reset price cap. In California‚ 10 percent is also added for credit risk. This Order made these price caps applicable for all spot transactions in the entire WSCC. Contracts that last for twenty-four hours or less define a spot transaction subject to the WSCCwide price caps. Thus‚ the June Order establishes the maximum spot price permitted in the California and west-wide spot market‚ absent justification‚ for non-reserve deficiency periods. That spot price (24 hours or less)‚ established during the 5

6

The actual Stage 1 emergency definition was 7 percent and FERC subsequently corrected this error. Order on Rehearing of Monitoring and Mitigation Plan for the California Wholesale Electric Markets‚ Establishing West-Wide Mitigation‚ and Establishing Settlement Conference‚ 95 FERC ¶61,418 (June 19‚ 2001).

162

The California Electricity Crisis: What‚ Why‚ and What’s Next

last reserve deficiency period on May 31‚ 2001‚ was $92 (plus 10 percent for sales within California). The new cap was to last until at least the spring of 2003. It was continued and remains in effect in 2004.7 On December 19‚ 2001‚8 about a year after the worst week in California’s energy crises‚ the FERC modified the west-wide price mitigation methodology to take into account likely higher natural gas prices in the winter. The new spot price cap established for non-reserve hours for the period December 20‚ 2001 through April 30‚ 2002 was $108 per MWH (plus 10 percent in California)‚ which was the mitigated spot market price set during the last reserve deficiency period.9

REFUNDS FERC’s initial attempts to place price caps on just California and then only during reserve deficiency or declared emergency periods proved futile. Several orders attempted to fix the problem. Ultimately‚ this was accomplished in June 2001. At that time‚ the FERC addressed its prior failures to control the market clearing prices by establishing potential ex post price refunds for the period October 2‚ 2000 through June 20‚ 2001. This period coincided with the FERC’s investigation into unjust and unreasonable spot prices and the period when its attempts to mitigate prices mostly failed. The full effects of California’s electricity crisis will take years to untangle and resolve. That said‚ the regulatory aftermath is coming into sharper focus as the FERC has ordered refunds to be paid for prices that exceeded what the FERC has defined as a mitigated market-clearing price (MMCP) for energy sold to the CPX and CAISO for the period October 2‚ 2000 through June 20‚ 2001. This refund period was established based upon the FERC’s statutory interpretation of when‚ based on the formal initiating of its price investigation‚ refunds could first commence. The word “refund” is a misnomer because it is doubtful sellers would actually ex post pay cash refunds to buyers because sellers have mostly not been paid for their sales made during the high-priced months of November and December 2000. The CPX and CAISO markets clear about three weeks after the close of the calendar month. The two larger IOUs were on the 7

The CAISO attempted to reduce the cap during a subsequent Stage 1 emergency. The CAISO was accused of gaming the process by falsely declaring an emergency so as to reset the cap. The FERC decided to restore the previous western state’s wholesale spot markets price cap. 8 Order Temporarily Modifying the West-Wide Price Mitigation Methodology‚ 97 FERC ¶61,294 (December 19‚ 2001). 9 The FERC determined to continue this price cap beyond April 30‚ 2002.

Refunds and Mitigation

163

brink of bankruptcy when the December bills were sent out. The state of California did not assume purchasing responsibility until January 2001. The result is that entering 2004‚ the sellers collectively have not been paid about $3 billion‚ based on the MCPs that were in effect prior to any FERC mitigation. The FERC’s refund proceedings has had two distinct stages: (1) the Administrative Law Judge (ALJ) Birchman and (2) FERC Staff proceedings. In the first stage‚ ALJ Birchman held hearings that established the regulatory aspects of creating a MMCP to replace the “unjust and unreasonable” MCP produced in the market during the refund period. The second stage occurred after the now-infamous and explosive Enron Memo and other admissions of corporate wrongdoing‚ including manipulation of the natural gas markets came to light. The FERC Staff proposed a second round of mitigation based upon using a different natural gas price than the reported spot market prices for natural gas market that ALJ Birchman recommended should be used in calculating the MMCP. FERC Staff proposed replacing these reported prices with a proxy price for spot natural gas using a formula based on the Henry Hub prices plus delivery costs to California. We address both approaches here.

ALJ Birchman’s Mitigation Plan The first step in refund liability determination is estimating an MMCP to replace MCP. The essential determinants of MMCP‚ or the competitive proxy price to replace the MCP for spot wholesale electricity prices‚ are based on the spot market price cap concepts discussed above. The key components of the MMCP are: Incremental heat rates for the CAISO’s real-time energy markets defined as the marginal units dispatched to meet the ISO’s real-time energy requirements for generators that provide the required heat rate data. (Effectively‚ all California natural gas-fired generation is included.) Daily reported spot prices for natural gas. $6 per MWH for operation and maintenance expenses. A 10 percent credit-adder for risk for energy sold after January 5‚ 2001 when SCE and PGE debt was downgraded.

164

The California Electricity Crisis: What‚ Why‚ and What’s Next The MMCP equals the highest northern (NP15)10 or southern (SP15) product of the incremental heat rate for the unit that cleared the real-time market and daily spot natural gas price‚ plus $6 for Operation and Maintenance costs (O&M). The MMCP is multiplied by one plus ten percent for post January 5‚ 2001 sales.

Several additional matters apply. Generators that pay NOx and other variable environmental compliance costs can claim these charges and submit proof of the expense. All sellers can voluntarily give up using MMCP in exchange for cost of service (COS) prices for all their spot market sales during the refund period. Sellers can reduce their potential refund liability by showing that applying the MMCP to their sales would result in a confiscatory result over their entire portfolio of sales into the California market. A very significant‚ highly disputed matter is how to use the MMCP to recreate the market during the nine refund months. Most sellers argued that if the market was to be re-created‚ the MMCP should replace the MCP every hour. Buyers and FERC Staff argued that the MMCP should be used as an ex post price cap. Thus‚ if MCP was less than the MMCP in any hour or transaction‚ the original MCP‚ not a higher MMCP‚ should apply. The economic arguments for simulating market outcomes would favor sellers. The regulatory argument that the FERC put sellers on notice that it might impose price caps favors buyers’ arguments. The matter has not fully been resolved as 2004 begins. However‚ the “lesser of” or price-cap approach seems more likely to prevail at the FERC than the replacement MMCP supported by the sellers. Judicial review had not been completed at the time this book was written. Regardless‚ Table 15-1 reflects ALJ Birchman’s findings on the effect of his refund recommendations‚ using the lesser-or rule to replace MCP with the MMCP.

10

NP15 and SP15 represent markets north and south of Path 15 in Central California. Different natural gas prices and heat rates could apply to both north and south. MMCP is equal to the greater of the two during each hour of the refund period.

Refunds and Mitigation

165

FERC Staff’s Approach After revelations concerning reported natural gas prices‚ the FERC Staff concluded:11 “ ...that published California natural gas price data are not sufficiently reliable to be used in the California refund proceeding for purposes of calculating the MMCP and resultant refunds.” The FERC Staff’s Report recommends using natural gas prices in the MMCP calculation that are significantly less than daily published natural modal prices that the CAISO proposed and Judge Birchman accepted. If the FERC Staff proxy prices are adopted for the Refund Period‚ refund obligations will increase significantly‚ effectively eliminating the outstanding pre-mitigation payments of about $3 billion for electricity. The principal reason given by FERC Staff for proposing to change the referenced natural gas price from California prices to external surrogate natural gas prices is that California’s natural gas prices were not highly correlated with or similar to natural gas prices in the primary eastern trading market at Henry Hub. Specifically‚ the Staff Report concludes: “The price data published in Gas Daily‚ NGI‚ and Inside FERC Gas Markets Report for the three California delivery points (Malin‚ PG&E Citygate‚ and Southern California Large Packages [parenthesis added]) should not be used for calculating the MMCP for the Refund Period...

11

Staff Report‚ Federal Energy Regulatory Commission‚ Initial Report on Company-Specific Separate Proceedings and Generic Revaluations; Published Natural Gas Price Data; and Enron Trading Strategies (Docket No. PA02-2-000) Fact-Finding Investigation of Potential Manipulation of Electric and Natural Gas Prices; August 2002. (Staff Report) See pages 56-57.

166

The California Electricity Crisis: What‚ Why‚ and What’s Next

“We are recommending...using producing basin spot prices plus transportation costs. “...Because natural gas at these producing areas (San Juan and Permian for Southern California and West Coast (Alberta) for Northern California [parenthesis added]) is actually delivered to California‚ Staff believes that this alternative is superior to the other alternative which Staff considered‚ the price at Henry Hub. “...given that prices at Henry Hub and at the producing areas highly correlate with each other. In contrast‚ the price indices currently being used for the California refund proceeding do not correlate well with Henry Hub prices.”12 The FERC Staff proposed to change the referenced natural gas price for determining MMCP to surrogate natural gas prices because FERC Staff speculated that traders may have manipulated the reported daily natural gas price data. The FERC Staff recommends using an average of San Juan‚ Permian‚ and West Coast prices. As evidenced by the quote above‚ the basis for FERC Staff’s contention was that the reported prices for California were not reliable and that the staff was unable to correlate the California prices with the prices that occurred at the Henry Hub during this period. The Report attaches great significance to the fact that natural gas prices for California were highly correlated to the prices at Henry Hub for the period prior to and the period after the price spikes‚ and that prices in other markets remained highly correlated to the prices at the Henry Hub during the Refund Period. Table 15-2 compares the original CAISO natural gas prices to the lower FERC Staff prices on a weekly basis. Starting in mid-November 2000 through the winter of 2000/2001‚ FERC Staff’s revisions were markedly less.

12

FERC Staff Report, p. 61.

Refunds and Mitigation

167

This proposed natural gas price reduced the corresponding estimated market clearing prices. This effect is shown in Table 15-3. Again‚ the late 2000 and early 2001 MMCPs are markedly less.

168

The California Electricity Crisis: What‚ Why‚ and What’s Next

Price spikes are not surprising during times when demand outstrips supply (caused by a combination of market forces including cold weather‚ drought conditions‚ low hydroelectric supplies‚ increased electricity related demand for natural gas‚ and a major pipeline explosion). It should also not be surprising that price spikes in California caused by localized market forces are not correlated with prices at Henry Hub.

Refunds and Mitigation

169

Chapter 11 discussed natural gas prices in great detail. As we explained‚ Henry Hub prices were highly significant as one explanatory factor in determining natural gas prices in California. Regional weather‚ economic factors‚ and other factors were also very significant. On an institutional level‚ pipeline capacity into California and “takeaway” capacity within the state is constrained during high demand periods. Competitive markets use price signals‚ such as price spikes‚ to signal that additional capacity is needed. In this context‚ competitive gas markets rely on price signals indicating the need for additional pipeline capacity. It is reasonable to believe that such price signals were part of the reason that the 900‚000 Mcf per day Kern River Pipeline expansion is being built. There also were similar reported price spikes (as high as $40 per MMBtu) at the Chicago city gate that were well above prices at the Henry Hub during several cold winters in the mid-1990s. These price signals helped spur the construction of the 1‚200‚000 Mcf per day Alliance Pipeline from Canada. High prices in the Northeast‚ again signaling deliverability tightness‚ continue to spur new pipelines and expansions. Such local market spot price spikes have little correlation to spot prices at Henry Hub when caused by distant‚ unique supply/demand conditions. Rather‚ an important function of local competitive market forces is to signal the basis for new investments. In short‚ prices are supposed to provide signals to market participants. The FERC Staff Report also questioned the quality of the pricing information provided by NGI‚ Gas Daily‚ and Inside FERC on the grounds that since the reported prices for California do not correlate with the prices reported for Henry Hub‚ these prices may have been manipulated. These referenced publications rely on surveys to develop the pricing indices that are relied upon by the industry—buyers and sellers alike. The only way these publications can obtain the information from market participants that is used to develop these price indices is to pledge confidentiality to the respondents. This does not mean that the information is unreliable. To the contrary‚ these publications strive to validate the pricing information they receive from both buyers and sellers‚ discarding prices they are unable to verify. These publications depend on their reputation to sell their indices. Furthermore‚ the FERC has relied on these indices in the past‚ and market participants and contracts continue to rely upon them today. The FERC Staff Report discusses the reliability of the information available for the California gas markets during the time period in question‚ again pointing with suspicion to the lack of correlation between prices in California and the prices at the Henry Hub. The Report‚ in suggesting that San Juan‚ Permian‚ and West Coast prices be used to calculate MMCPs during the Refund Period‚ points with satisfaction to the fact that the correlation between prices at San Juan/Permian with prices at Henry Hub

170

The California Electricity Crisis: What‚ Why‚ and What’s Next

remained high during these critical few months in late 2000 and early 2001. Thus‚ the FERC Staff Report concluded that the prices (albeit from the same publications‚ developed using the same methodologies that it questions for California prices) for San Juan‚ Permian‚ and West Coast should be used to calculate MMCPs for the Refund Period. However‚ the FERC Staff Report failed to provide any analysis about whether prices in California actually spiked to the levels reported in the publications referred to above and thus‚ whether the lower surrogate natural gas price recommended by the FERC Staff accurately reflects the prices paid for natural gas. The fact that natural gas spot prices spiked in California is reliably reported in many publications. The CEC reports that prices spiked at around $60 per MMBtu during the Refund Period.13 That high prices were actually paid in the natural gas spot market is a fact that is undisputed. The FERC Staff Report provided no analysis that would explain why prices spiked in California. The facts are widely known and virtually undisputed. For example‚ in October 2001‚ state analysts at the CEC issued a report entitled Natural Gas Infrastructure Issues (CEC Report) that examines in detail the factors that caused the price spike in natural gas prices experienced in California in late 2000 and early 2001. We discussed these factors in great detail in Chapter 10. As has been recognized for electricity markets‚ there can be localized (fairly large in fact) constrained markets for natural gas. Behind the constraints‚ market swings in prices are likely to be more volatile because these constrained natural gas markets respond to both worldwide petroleum conditions plus localized swings in demand and supply. The FERC Staff Report grossly fails to grasp this fundamental fact. In effect‚ FERC Staff seeks to ignore the factual circumstances in the west during this period‚ including severe climate in late 2000 that combined drought and a cold late fall/early winter‚ a lack of sufficient summer storage fills‚ constrained pipeline capacity within the state (usage in excess of 100 percent utilization)‚ and a phenomenal surge in electric system demand for natural gas. The key point is that these two markets were both severely affected by a combination of market forces and infrastructure problems that were not faced by other markets in the country. There is no reason to think that natural gas prices would not spike in an area where demand severely outstrips supply. There is no reason to think that prices in an area affected by market forces‚ which cause prices to spike‚ would be highly correlated to prices in an area not affected by those market forces. In fact‚ there is good reason to suspect just the opposite‚ that the prices in the two separate and diverse markets would not be highly correlated. 13

CEC Report‚ p. 69 and Cato Report‚ p. 5.

Refunds and Mitigation

171

Accordingly‚ there is no valid reason to think that prices reported for markets (i.e.‚ San Juan‚ Permian‚ and West Coast) that were not subjected to these market forces and infrastructure restrictions would have any relationship to the prices reported for markets that were subjected to these market forces (i.e.‚ California). The fact that prior to and subsequent to the demand conditions that existed in the California markets‚ prices in California correlated to prices at the Henry Hub is not surprising. Nor is it surprising that during a price spike caused by circumstances peculiar to a particular market‚ the prices would diverge from other markets. On a statistical level‚ natural gas prices in late 2000 were actually significantly higher than the “normal” differences between California and Henry Hub would predict. Statistical and institutional analyses account for most of the price differences. There is also extraordinarily high forecast accuracy as we discussed in Chapter 10. Data manipulation has not been proven. The proxy price for natural gas approach proposed by FERC Staff also effectively means that California would not have to pay about $3 billion in unpaid electricity bills for late 2000. When we wrote this book‚ the final outcome was still to be written‚ and it is not at all certain that there will be a final accounting any time soon.

This page intentionally left blank

Chapter 16 CALIFORNIA RESPONDS

In 2003‚ California recalled Governor Davis. Two factors were highest on the list of reasons for his recall. First‚ there was the state’s indebtedness caused in part by the state’s assumption of electricity purchasing. Second‚ the recently re-elected Governor was still being blamed for the state’s energy crisis‚ which coincided with economic slowdown. Prior to the recall election‚ California’s principal energy agencies joined ranks to create a unique document. The CPUC‚ the Consumer Power and Conservation Financing Authority‚ and the Energy Resources Conservation Development Commission have all approved and released their Energy Action Plan. The three entities state that the Energy Action Plan is intended to be a blueprint‚ flexible enough to change depending on circumstances that will support the state’s economic growth and attract new investment. The Energy Action Plan identified six specific goals that should be undertaken to ensure that California continues to grow and prosper. These are: Meet California’s energy growth needs while optimizing energy conservation and resource efficiency and reducing per-capita electricity demand. Ensure reliable‚ affordable‚ and high quality power supply for all who need it in all regions of the state by building sufficient new generation. Accelerate the state’s goal for renewable resource generation to 2010. Upgrade and expand the electricity transmission and distribution infrastructure and reduce the time before needed facilities are brought online. Promote customer and utility-owned distributed generation. Ensure reliable supply of reasonably priced natural gas.

174

The California Electricity Crisis: What‚ Why‚ and What’s Next

While necessary‚ this Energy Action Plan may not be sufficient. The power crisis in California has caused a political reaction at least as significant as the crisis. This manifests itself in calls for change from both the left and the right. Some demand change that eschews deregulation‚ blames greedy businesses‚ and seeks to punish unjust gaming. Others blame politicians and regulators for not letting the market work or not designing it properly. California’s consumers and voters demand a new system that will protect them from extreme prices increases‚ volatility‚ and politically dictated and seemingly unfair cost assignments. California has previously taken a pragmatic approach to electricity and what entity will supply this vital commodity. Indeed‚ in the early twentieth century‚ there was a long debate over whether power should be publicly or privately supplied. The history surrounding the early development of Los Angeles and the construction of the Hoover Dam‚ the Los Angeles Aqueduct‚ and the formation of what is now the Los Angeles Department of Water and Power provides evidence of the state’s pragmatic approach. By using a combination of private business and public money‚ Los Angeles secured solutions intended to produce the best possible reliability result and lowest electricity prices it could achieve. Differences in political philosophies yielded to least-cost and can-do sentiments. Thus‚ California has historically supported those pragmatic approaches‚ whether they were public‚ private‚ or a combination that worked the best. It is impossible to be certain about the future. The joint agency Energy Action Plan presents a possible beginning. This chapter discusses some proposals that would establish a new public/private partnership electricity industry for California. The IOUs are also seeking to emerge from the depths of their downward slide. It is too early to know what will happen next with certainty. California needs to expand its energy and related infrastructure at an affordable cost. A legislative and popular decision to expand the role of municipalities‚ either on their own or in direct partnership with the IOUs‚ would be a pragmatic‚ not philosophical‚ shift for California. Such a future would fall within California’s historic tradition that solves problems efficiently while avoiding labels and political philosophy. It is possible‚ but not likely‚ to imagine a future in which California’s IOUs regain their financial strength‚ restore their business and marketing verve‚ and once again gain consumer and political confidence. It could be years before such conditions would return to California. As 2004 begins‚ two events are important: SCE’s plans to build and own generation and PG&E’s imminent emergence from bankruptcy. Nevertheless‚ given the things that would still have to change for IOUs to take roles similar to those they generally admirably performed in the last century‚ it does not appear

California Responds

175

realistic at the start of the twenty-first century to think‚ especially with a new governor who supports competition‚ that California’s IOU-centered electricity establishment will recover in the immediate future. The tasks facing the IOUs are still daunting. The IOUs have gained strength. However‚ the lOUs are still associated with many of the things that went bad starting in 2000. It still seems politically unlikely for California to reward the utilities (i.e.‚ to restore their financial health) in the near term so that they would have the financial wherewithal to make necessary investments. Competitive retail Energy Service Providers (ESPs) could also potentially evolve in California’s future. ESPs appeared ready to expand significantly in California before the power crisis hit. As with the IOUs‚ the ESPs are currently perceived to represent potentially costly and all too risky non-regulated options. Again‚ this could all change. But for now‚ neither IOUs nor ESPs working alone (i.e.‚ without partnering with municipalities) seem likely to have the necessary consumer or political support to be able to succeed. The state’s population is expanding. Most important‚ the state’s economy and population currently need energy infrastructure and new power sources. Waiting for utility financial viability to return does not seem to be an option. Outsiders still consider California to be “golden” in terms of climate‚ resources‚ and opportunities. This approach puts additional pressure on the failing‚ existing infrastructure. In addition‚ there is a continuing need to add jobs and places for people to work and conduct commerce. There are several opportunities in California for much needed growth. These are mostly being stymied by the current financial problems faced by the state’s utilities and the IOU’s financial difficulties will continue for at least the intermediate future. Utility distribution is one part of the necessary infrastructure for new industrial parks‚ business complexes‚ and shopping centers. First‚ and foremost‚ California must spend money to enable this distribution or wiresrelated growth. The IOUs definitely seek to be involved. In fact‚ SCE has instituted an extensive advertising campaign to assert its role in this area. However‚ the IOUs have not yet fully recovered financially. Furthermore‚ the state intends to shift the liability for its long-term energy contracts to the IOUs. If other financing were available‚ the IOUs would not necessarily need to own or invest in this needed new energy distribution infrastructure. The dilemma is that distribution is where retail consumers attach to the grid. The IOUs do not want to lose customers to new municipal utilities or hybrids such as community choice aggregators (CCAs). The prospect of losing

176

The California Electricity Crisis: What‚ Why‚ and What’s Next

customers to ESPs is less because these entities are‚ as 2004 begins‚ not even authorized to add customers in California.

NEW MUNICIPAL UTILITIES: SEVERAL PUBLIC/PRIVATE OPTIONS ARE POSSIBLE FOR CALIFORNIA Municipally Owned Utilities (MOUs) Each municipality in California has the opportunity to form a Municipal Power Authority (MPA)‚ which would have the power to issue bonds‚ purchase assets‚ make infrastructure improvements‚ and execute energy service contracts. Each MPA could also be part of a larger umbrella organization that would encompass other regional municipalities. The MPA (or the umbrella group) would be authorized by the respective city councils to negotiate with the incumbent IOU to purchase the existing IOU’s distribution system under a new partnership agreement with the IOU. An alternative approach is less ambitious and would simply be for the MPA to take over new or incremental undeveloped portions of the IOU’s franchise within the boundaries of the municipality or adjacent to it through annexation. Under this alternative‚ the IOU could retain (or sell to a municipality) all or part of the pre-existing IOU distribution system. The MPA would‚ however‚ assume responsibility to build and own the new systems to serve new areas‚ particularly new business enterprise zones as a MOU. MOUs currently serve two of the larger cities‚ Los Angeles and Sacramento. Several other California cities also have MOUs.1 Additional cities are studying the possibility of or proposing to form new MOUs. The IOUs generally strongly oppose these efforts. MOUs would expect to pay a premium over depreciated book value to acquire IOU assets. The municipalities would finance these investments with tax-exempt bonds. Thus‚ despite a premium over depreciated book value‚ the net cost to the municipality could be lower than that of the IOU due to both the municipality’s increased leverage and lower cost of money. Similarly‚ new or incremental infrastructure investment would be financed using the municipality’s favorable financing terms and superior access to

1

For example‚ Alameda‚ Anaheim‚ Azusa‚ Banning, Biggs‚ Burbank‚ Colton‚ Escondido‚ Glendale‚ Gridley‚ Heraldsburg‚ Lodi‚ Lompoc‚ Needles‚ Palo Alto‚ Pasadena, Redding‚ Riverside‚ Roseville‚ Sand Francisco‚ Ukiah‚ and Vernon.

California Responds

177

capital markets. These new business enterprise zones would likely cost less‚ and MOUs would pass some infrastructure costs on to developers.

A New Servco With Joint Ownership Concurrently‚ utility holding companies could form a service company subsidiary (Servco)‚ which would not be subject to CPUC retail price and cost of service regulation. SCE already has such a non-regulated service company that provides operations and maintenance service for other generation-owning entities (e.g.‚ small municipal utilities) that outsource these functions because they prefer to rely on SCE’s expertise. A similar type of service company could be formed to provide operations and maintenance services for newly formed distribution services that MOUs provide. The Servco could add other management and retail services. These new utility-owned Servcos would not fall under the CPUC’s regulatory auspices. Rather‚ they would become fully non-regulated management services and operational companies freed from regulatory restrictions and subject solely to contracts freely entered into with the municipalities. A particularly important Servco function would be jointly developing new municipal enterprise zones with MPAs that would be owned primarily by municipalities and financed by them under favorable tax and interest terms. The IOUs’ non-regulated affiliate would design‚ operate‚ integrate‚ and maintain these new enterprise zones and‚ perhaps‚ jointly own these non-CPUC-regulated enterprise zone and new municipal utility systems. Services to customers would likely include electricity delivery services‚ new fiber-optic services as well as energy service contracts that would provide various types of price/risk retail service options. The utility’s investors would benefit from their Servco ownership. Contract negotiations with MOUs could cause acquisition premiums to be reduced. There are‚ however‚ regulatory or legal impediments to MOU formation that affect matters in California.

Community Choice Aggregators (CCA) California has enacted a statute‚ AB 117‚ that provides for forming a new entity in California known as a CCA. These entities would assume responsibility for acquiring or purchasing generation (MWHs). The IOUs would deliver CCA provided electricity to retail consumers that receive CCA service. The IOUs would continue to own and operate the distribution‚ metering‚ and perhaps the billing services for CCA consumers. In effect‚ CCAs are municipally owned electric service providers (ESPs). CCAs are the only permitted new competitive retail choice provider

178

The California Electricity Crisis: What‚ Why‚ and What’s Next

permitted in California as 2004 begins. Although AB 117 authorizes CCAs‚ it also requires some payment of exit fees for retail consumers that switch from IOU to CCA service. Similar restrictions would also apply to new MOUs that form where existing IOU franchise services are currently provided. The IOUs have averred and the CPUC has held that exit fees should also apply to new Greenfield‚ or undeveloped‚ MOU service territories.2 Imposing exit fees on greenfields is a particularly important issue for several municipalities that seek to develop closed former military bases in order to build their local economies and expand their cities.

Joint Powers Authorities (JPA) The municipalities in California have the authority to jointly build‚ own‚ and operate new generation. These JPA generating stations benefit from open transmission access. Independent power producers (IPPs) or IOUs could partner with the MOUs and the JPA. The municipality would assume responsibility for guaranteeing the customer base and would‚ at least initially and at its discretion‚ assume most financing and ownership responsibilities. In conjunction with advice from the Servco‚ the MOUs that participated in the JPA would be responsible for all additions‚ expansions‚ and improvements in the distribution system‚ especially in new enterprise zones. The MOU would become vertically integrated. The wholesale competitive market for spot trades would continue. Long-term purchase power contracts‚ JPA construction‚ and other ownership could provide generation. The CAISO would provide transmission. MOUs‚ Servcos‚ and CCAs would provide distribution and metering services. There would be some migration from privately owned IOU distribution systems to MOUs. To the extent pre-existing distribution assets are in play‚ this transformation would be negotiated by incumbent utilities and the municipalities. Competitive entry could be restricted to new enterprise zones. These areas likely would be appended to the service agreements that are negotiated between the adjacent IOUs and the cities. To the extent that Servcos are not formed‚ the municipalities would not need to negotiate exclusively with the incumbent IOUs to create new enterprise zone Servcos. The municipalities would be free to develop these on their own or in partnership with other entities.

2

As of the date this book was written‚ the California Supreme Court was considering an application to hear an appeal of the CPUC decision.

California Responds

179

New Generation Capacity The key downstream issues that flow from the wholesale power grid are generation‚ distribution‚ and what entity serves retail consumers. Upstream‚ California needs additional electric generation capacity. Neither the IOUs nor Independent Power Producers (IPPs) have been willing or able to assume the political risks inherent in building new generation in California. In 2000‚ Governor Davis promised that 20‚000 MWs of new generation would be built in California by 2005. This promise has remained largely unfulfilled. Only about 5‚700 MWs of this new generation is now operating. There are 6‚635 MWs of new generation with approved licensing. Construction on about half (3‚300 MWs) the licensed generation has been suspended or cancelled. Further‚ according to the California Energy Commission‚ about 5‚000 MWs of new capacity has been withdrawn from the power plant siting process.3 Conservation efforts have helped to reduce demand in California in 2001 and helped end the electricity crisis. Regardless‚ California’s population is pushing toward forty million. This population growth will increase electricity consumption. For example‚ peak electricity demand in Southern California alone‚ according to the Southern California Association of Governments (SCAG)‚4 is expected to increase by 29 percent in 2012.5 This projected increase in demand is shown for SCAG alone in Table 16-1.

3

Arthur O’Donnell‚ “Sowing the Seeds for California Crisis II‚”Public Utilities Fortnightly‚ (May 1‚ 2003)‚ p. 37. 4 Southern California Association of Governments (SCAG)‚ Regional Comprehensive Plan and Guide–Energy (Chapter Update 2002). 5 This does not include SDG&E’s current peak electricity demand or any growth for SDG&E.

180

The California Electricity Crisis: What‚ Why‚ and What’s Next

In early 2003‚ the IOUs did not have the financial wherewithal to build the generation necessary to meet this projected increased demand on their own. The IOUs are recovering financially. Their stock prices are increasing. In late 2003‚ PG&E received a CPUC retail rate order that will pull it out of bankruptcy. All PG&E’s creditors will be repaid. The retail consumers will pay about $8 billion in surcharges above their current cost of service (COS) over the next decade. The CPUC has also virtually guaranteed investors and lenders that PG&E will earn its authorized ROE over this period. SCE has also been recovering financially in 2003. SCE avoided bankruptcy by working with the state’s CPUC and Governor. SCE received a favorable rate case that provides for SCE to collect future money to compensate its shareholders and the state for MWHs purchased and longterm contracts signed during the energy crisis in 2000 and 2001. The political and economic problem is that these IOUs had prices twice the national and regional average in 1996 when AB 1890 was enacted to restructure California’s electricity industry. The IOU’s retail rates now likely to apply in California for the next decade are close to three times the

California Responds

181

rates in other parts of the United States despite recent modest rate reductions. With high retail prices‚ the IOUs have more than a tarnished image. Most consumers and voters are angry. To retain consumers‚ the IOUs convinced the legislature‚ and are now working to convince the CPUC‚ that current‚ former‚ and potential new IOU customers must pay to restore financial order‚ whether they remain IOU customers or depart to be served by JPAs‚ CCAs‚ or MOUs. The IOUs argue that these policies are fair. Regardless‚ these policies also undermine‚ but do not eliminate‚ the value that new MOUs‚ CCAs‚ and ESPs might provide retail customers. Much is at stake. This is a work in progress. This saga in California also involves new generation construction. With financial health returning‚ the IOUs seem eager to insure their future dominance. To this end‚ SCE has proposed to build and own a new 1054 MW generation facility‚ the MountainView facility (MTV)‚ in Redlands‚ California. This SCE-owned facility would be unique. MTV would not become part of SCE’s utility-owned rate base. Instead‚ SCE would purchase power from itself under a FERC-approved contract known as a Purchase Power Agreement (PPA). The PPA would move MTV from CPUC to FERC regulation. The FERC regulation would be light-handed wholesale generator PPA regulation‚ not FERC comprehensive COS regulation. The year 2004 began with the CPUC approving MTV and its PPA‚ and an SCE submission to FERC for speedy approval. Consumer groups‚ IPPs‚ staff members at the CPUC‚ and others are strongly opposed to this special arrangement. Two things are apparent. First‚ the IOUs are re-emerging. Second‚ new municipal entities predicated on a need to fill the IOU void are threatened before they fully engage. The IOUs are still way behind due to the blame attached to the crisis and the constant reminder provided by high retail prices. If the MOUs‚ CCAs‚ and JPAs are to succeed‚ they need regulatory‚ legislative‚ judicial‚ and political help so they can achieve lower prices than the IOUs. Municipalities can also build generation‚ enter long-term power supply contracts‚ and enter other power deals where others build‚ own‚ and operate new power stations under JPAs. The new MOU/Servco joint ventures could also jointly construct and own new generation. Joint-venture service contracts could also be executed to supply light‚ heating/cooling‚ and power to municipal and school buildings. The success of municipalities in the electricity industry is uncertain in this battle for California’s future electricity business that began in earnest in late 2003.

182

The California Electricity Crisis: What‚ Why‚ and What’s Next

Distributed generation6 and renewable energy are often economic and could be used to meet a portion of the projected increased peak demand‚ thus delaying or trumping the need to build new combined-cycle combustion turbines (CCCTs) or other generation plants. In this way‚ distributed energy can help increase reliability and security. Also‚ to the extent that distributed energy uses renewable energy sources like solar‚ wind‚ biomass‚ or fuel cell technology‚ it can also provide environmental benefits. Municipally owned utilities often have the financial resources and commitment necessary to acquire these new and now much improved technologies. The IOUs have some significant experience and expertise in these matters. Nevertheless‚ their ability to expand new renewable and environmentally friendly technologies is currently limited. A hybrid approach could combine the economic strength of the municipalities with the expertise of the IOUs to explore and develop these new technologies for California and overcome some of the regulatory barriers that have slowed distributed energy’s use in California.7 Alternatively‚ the battle for dominance could distract good intentions and reduce successes. Demand side management (DSM) programs can also reduce energy consumption and demand. DSM programs fall into two broad categories: (1) demand management and (2) energy efficiency. A well-known example of demand management is Governor Davis’s 20/20 program‚ where consumers who reduced their electricity consumption by 20 percent received a 20 percent reduction in their electricity bill. Energy efficiency refers to energy saving devices and appliances. DSM programs have been successful in the past. There has‚ however‚ always been an inherent conflict between IOUs (which are in the business of selling kWHs) and DSM programs (which endeavor to reduce the kWHs consumed). No other business attempts to get its customers to consume less of the product it sells‚ thereby reducing its revenue. This conflict was exacerbated with electricity restructuring because the CPUC no longer sets IOU generation-related profits‚ further reducing the IOU’s incentives to get customers to reduce their consumption. After deregulation‚ the price freeze eliminated price signals‚ at least initially‚ that would have provided consumer incentives to reduce demand or conserve based on strictly economic motivations. 6

7

The California Energy Commission (CEC) has defined distributed energy as “electric generation connected to the distribution level of the transmission and distribution grid usually located at or near the intended place of use.” CEC‚ Distributed Generation Strategic Plan‚ June 2002‚ P700-002-002‚ p. 2. The CEC reports that there are currently 500 distributed energy facilities within SCE's service territory producing 766 MW (Ibid.‚ pp. 8-9). Barriers to entry include “high fees‚ long approval processes‚ insurance requirements‚ exit fees‚ and capital costs.”

California Responds

183

To help pay for energy efficiency construction programs‚ a public goods charge8 is collected by IOUs‚ through tariff surcharges. About $300 million per year is allocated for this purpose.9 The CPUC is currently setting aside about 20 percent of this amount annually to be administered by non-utility parties. These parties include cities‚ counties‚ and governmental councils. These dollars could help provide investment for MOUs to implement new‚ effective DSM programs. Unlike the IOUs‚ which often have a built-in conflict in implementing DSM programs that affect their revenue‚ MOUs have no such built-in conflict. MOUs owe a duty to the residents of their municipality‚ not to shareholders‚ and would find it less difficult to subsidize reduced energy sales. The benefits of a successful municipal DSM program will positively affect residents by improving efficiency‚ both end use and system supply‚ and DSM will also help keep prices low by delaying or deferring new construction profits. The interests of the MOU and the municipality’s residents would mostly be aligned. The California Power Authority (CPA)‚ a relatively new entity arising from the state’s energy crisis‚ recently adopted its Energy Resources Investment Plan. Under that plan‚ the CPA can issue up to $5 billion in revenue bonds that can be used to push electricity markets toward cleaner‚ more efficient energy and greener‚ renewable resources. On April 10‚ 2003‚ the CPA announced its first bond sale of $28 million to provide “energy conservation funding to schools‚ cities‚ counties‚ non-profit/public hospitals and public care institutions in California.” (10) In the future‚ the CPA revenue bonds could also assist MPAs in funding conservation and renewable energy development.

8

See CPUC 2001 Energy Efficiency and Conservation Programs‚ Report to the Legislature, Prepared by the Energy Division‚ December 2001. 9 See Restoring California’s Electricity Future: The Rebirth of Utilities’ Resource Portfolio Management Responsibilities‚ [Background Paper for Senator Byron Sher]‚ Ralph Cavanaugh (March 23‚ 2003). In that paper‚ Mr. Cavanaugh describes a key element of the legislative plan to restore California’s electricity future as SB 1038‚ “which authorizes the renewable-energy funding needed to execute the SB 1078 mandate. The bill allocates some $1 billion across a variety of clean-energy programs from 2002-2007‚ all of it funded through the system benefits charge on electric rates. Of this‚ some $70 million per year ($350 million total) is available to cover costs of meeting the renewable energy portfolio standard obligation. See Public Utilities Code §383.5(d). These funds may be used to support renewable energy systems anywhere within the Western Electricity Coordinating Council’s transmission system–spanning the eleven Western States‚ Northern Mexico‚ Alberta‚ and British Columbia–as long as the seller has guaranteed contracts to sell its generation to end use customers’ of the California investor-owned utilities.” 10 State of California Consumer Power and Conservation Financing Authority Press Release (April 1‚ 2003).

184

The California Electricity Crisis: What‚ Why‚ and What’s Next

CONCLUSIO N California has choices. The worst seems to be in the past. The costs of recovering from the 18-month energy crisis are high. These costs will be spread over about a decade. There is great uncertainty over who or what types of institutions will emerge as the dominant suppliers in the future. Cooperation could yield benefits under new partnerships. The cities could play an expanded role. The IOUs are returning. It is impossible to predict the future. A mix of competition and regulation is certain. Who will survive and who will emerge is less certain. The next chapter addresses some of this in greater detail.

Chapter 17 HANDICAPPING THE WINNERS

The future structure of California’s electricity industry depends largely upon how the courts and regulators assign the costs needed to pay for California’s electricity crisis. If FERC adds more money for refunds, California’s retail consumers benefit. If the California economy expands, more MWHs sold would reduce the future per unit surcharge necessary to pay down the costs of the electricity crisis. A very important determining factor for the structure and ownership is how much of these costs new municipal entities would be assigned for the electricity crisis of 2000 and 2001. A second important matter is how much IOUs may be favored if they acquire generation. Related to this is to what extent IPPs and JPAs will be discouraged from building new generation. This chapter discusses some of the salient aspects of these debates. The core issue in much of what will happen in California is the Recovery of Utility Costs portion of the Order Instituting Rulemaking to Implement Portions of AB 117 Concerning Community Choice Aggregation. 1 As the Commission noted in its Order, AB 117 requires that the CPUC establish cost responsibility surcharges (CRS) prior to the time that a Community Choice Aggregator may begin aggregating load. The CPUC stated that “these surcharges allow the utility to recover certain energy purchase costs the utility would continue to incur after losing customers to the CCA.” The California Legislature, in enacting AB 117 was concerned that if customers departed the IOUs in favor of the CCAs, the remaining IOU customers would be stuck with the liability for the costs originally incurred on behalf of the customer base that included the customers who departed the IOUs for the CCAs.2 1 2

Rulemaking 03-10-003. See AB 117, Section 366.2(d)(e)(f).

186

The California Electricity Crisis: What, Why, and What’s Next

In its 2002 Order referenced above, the CPUC proposed to apply to the CCAs the same CRS and price cap that it adopted for ESPs.3 The CPUC majority reasoned that since CCAs and ESPs both provide electric service, both rely on the IOUs’ distribution and transmission services, and both CRS would include identical cost components,4 it was reasonable to set the CRS for CCAs equal to the Direct Access (DA) CRS currently applicable to ESPs. Thus, the CPUC proposed to apply a CCA CRS of $0.027 per kWH. In addition, the CPUC stated that it would apply any future changes in the DA CRS to the CCA CRS. In its Order, the CPUC also stated that it would consider parties’ proposals to reduce the CRS in certain circumstances. However, the only example provided by the CPUC was “where a CCA assumes liability of a utility’s CDWR energy contract commitments.” The Order does not consider that there are other benefits for California and remaining IOU customers if and when CCA supplied customers depart IOU service. The CPUC discusses the issue of a CRS on Municipal Departing Load (MDL) in the context of its Order Instituting Rulemaking Regarding the Implementation of the Suspension of Direct Access Pursuant to Assembly Bill 1X and Decision 07-09-060.5 In that decision, the CPUC stated that the Legislature’s stated intent in AB 117 was “to prevent any shifting of recoverable costs between customers.” 6 Thus, the CPUC concluded that “the potential for cost shifting is not limited just to DA customers, but also implicates other load that departs from IOU service, including customers that depart bundled service after February 1, 2001 to be served by a municipality.” While the question of departing load that had been served by the IOUs prior to February 1, 2001 was relatively clear cut, there was an issue with respect to whether new municipal load that had never been served by the IOUs also fell under the parameters of AB 117 and its requirements for imposing a CRS. In a 3-2 decision, the CPUC concluded that new municipal load would also be required to pay a CRS, reasoning that although these new customers may not have been served by the IOUs, the load these new customers represented had been considered in the CDWR procurement decision. Thus, to the extent that new municipal load had been included in projected load, 3

See Docket No. D. 02-1-022. The cost components are CDWR purchase costs, bond-related and administrative costs for CDWR purchases, and other unrecovered costs. 5 See for example Order Granting Limited Rehearings on the Issue of the Allocation of the Exception for New Municipal Load, Modifying Decision (D.) 03-07-028 For Purpose of Clarification, and Denying Rehearing of Decision, as Modified, in all Other Respects, Decision 03-08-076, August 28, 2003. 6 See AB 117, Section 366.2(d). 4

Handicapping the Winners

187

the contracts secured by CDWR to serve those projected needs were procured to serve that new municipal load. Therefore, the CPUC found that new municipal load was also responsible for paying the CRS even though this new load had never actually been served by the IOUs. The CPUC explained in its Order that there should be some limited exemption provided to new load and granted a limited rehearing to consider “whether, or to what extent, there is sufficient factual basis for CRS allocation based on whether the publicly (governmentally) owned utility was formed before or after February 1, 2001.” The CPUC identified five issues that needed to be addressed in order to ascertain whether there should be any exemptions for new municipal load and to sort through allocation issues between existing publicly owned utilities and to newly formed publicly owned utilities. Nevertheless, pending the hearing’s outcome, all departing pre-existing, as well as new municipal load, will be responsible for paying the CRS. How these issues are determined will go a great way toward defining the winners and survivors in the California electricity industry’s future. However, there is an important principle in regulatory economics concerning discriminatory treatment. While it is admirable that the Legislature is concerned that AB 117 and the formation of CCAs does not result in unreasonable cost-shifting to those customers who remain with the IOUs, it is also important to be pragmatic and consider the total effect that CCAs will have on the prices that the remaining customers will pay. Regulators often recognize that price discrimination, or allocating different costs and prices for services to similar customers, can be just and reasonable in certain circumstances. This is sometimes called “due discrimination.” The essential condition of any justified price discrimination is that both the “favored” customers and the “discriminated against” customers need to be made better off. To understand how this may happen, it is necessary to realize that not all cost allocations are “zero sum games” in which: (1) there is a fixed amount of pain or cost recovery to share; and (2) that there are not significant effects or consumer responses to the cost allocation process. A classic case supporting “due discrimination” would be a cost allocation where a full or equal cost allocation would cause some pre-existing customers to: (1) shut down (e.g., a marginal business); (2) move elsewhere (e.g., relocate to a more favorable economic climate); and/or (3) cause potential new customers to decide not to move to a particular place. In any of these cases, the total dollars paid by the “others” that remain are effectively increased as they bear the burden of the lost pre-existing customers and the foregone opportunity of some reductions if newcomers pick up some of the cost recovery. Regulators have recognized in such circumstances that partial cost allocation, thus a form of discrimination away

188

The California Electricity Crisis: What, Why, and What’s Next

from full cost allocation, is in the public interest and represents a “win/win” outcome. First, the favored customers win because they pay less. Second, the “others” also would pay less than they would otherwise have paid because their cost responsibility has been eased in that customers that might otherwise shut down, move away, or not move in pay some partial costs recovery. Third, departing customers may produce other direct benefits for the remaining customers. In addition to retail prices, there is also a second broad justification for “due discrimination.” This comes in the form of an array of potentially significant public goods or benefits. Discrimination is often justified if customers who pay less for a particular cost allocation build or acquire things of significant benefit (e.g., build a renewable energy generating station, expand energy efficiency or conservation efforts, relocate to less pollution-sensitive areas, reduce local unemployment, etc.). The CPUC will likely consider the proper role of “due discrimination” as it sets the CRS for CCAs, much as it has for conservation and renewables. The CPUC can, in effect, create a win-win situation where the overall prices paid by IOU customers will decrease as a result of customers choosing CCAs, even if the CRS for the CCA’s customers, especially new load previously not served by the IOUs, is set lower than the CRS for direct access customers previously set by the CPUC’s recent Order. A threshold question deals with how the CCA will secure the power needed to serve its customers. There are three potential available options for the CCA: (1) secure power from the same wholesale power markets utilized by the IOUs to serve their customers (e.g., drink from the same well as the IOUs); (2) secure a long-term power supply contract from sources outside the wholesale power market in which the IOUs currently purchase power; or (3) build a power plant from which the CCA’s customers can be served. On first impression, the first option produces no discernable potential benefits to California as a whole. The distribution and transmission function would still be provided by the IOU and the generation source would still be the same, albeit the purchasing party would be different. Regardless, the other two options do provide potential benefits to the state and to the customers remaining with the IOUs because load is reduced and new supplies are acquired. These potential benefits alone support a reduced CRS for the CCAs. Assume that a newly formed CCA takes 500 MW of the load from the IOUs’ load (which is, let’s assume, approximately 40,000 MW in California). Rather than relying on the same sources of generation that the IOUs did to provide this generation, the CCA would either secure alternate long-term supply sources from sources outside the wholesale market used by the IOUs or would (as a government utility) construct its own 500 MW

Handicapping the Winners

189

generating plant. The effect of this is to reduce demand in the wholesale market by the 500 MWs (assuming no increase in load). This would, all other things being equal, cause the price in the market in which IOUs purchase more than two thirds of their energy for resale to be reduced. Eliminating 500 MW out of a total IOU wholesale market of about 40,000 MWs (including utility-retained generation and the CDWR contracts) reduces total demand by about 1 to 2 percent.7 This shift may not seem large for the electricity market in California, but it likely would have a small effect on the price per MWH. When this small price effect is multiplied by the total number of MWHs the IOUs sell in California, the effect in terms of dollars saved by “others” can be become very large. These savings will all accrue to the benefit of those customers that remain IOUs’ customers. This large dollar savings can be compared to the value of the CRS that would be extracted from those customers who departed the IOUs’ system in favor of the new CCA. Again, let’s assume a 500 MW plant with an 85 percent capacity factor. That plant would produce 3,723,000 MWHs (500 * 8,760 * .85) per year. Some of this output would be used directly to supply CCA load. The remaining output would be used to supply surplus energy to the grid or wholesale market. Regardless, wholesale prices would likely decline for “others” not served by the CCA. For illustrative purposes, we will use the $2.70 per MWH CRS established by the Commission for direct access customers. Multiplied by the $2.70 per MWH CRS, the nearly four million MWHs would result in an annual CRS payment upwards of $10,052,100, assuming all the MWHs were used by the CCA’s retail customers. If the CRS payment were cut in half for departing load that went to a CCA that built its own generating plant, the CRS payment would be about $5,000,000. In order to justify the unequal CRS payment, it would be necessary to show that those customers who remain with the IOUs would pay less in the future. Recall that we assumed that total IOU-related demand in California was 40,000 MW, and for simplicity let’s call this 250 million MWHs per year. In order to achieve a $5 million savings over the 250 million MWHs, it would be necessary that the effect of reducing demand across the wholesale market reduce price by $0.02 per MWH, or .002¢ per kWH, a miniscule price reduction. When multiplied across the great number of MWHs sold each year in California, even such a tiny price reduction matches the reduced CRS paid in the above example. Stated another way, if the reduced demand on the system reduced wholesale market prices by 7

And if the CCA overbuilds this generation plant, it should be able to supply the excess power to the wholesale market, thereby increasing the supply and, all other things equal, decreasing the price.

190

The California Electricity Crisis: What, Why, and What’s Next

pennies per MWH, the savings across the entire system could easily exceed any reduction in the CRS paid by the CCA customers. Mathematics aside, the basic case is how much “others” still served by IOUs would benefit from JPAs building needed capacity for California. Given the financial challenges faced by IOUs, such additional investments, regardless of the accuracy of the above mathematics, would benefit remaining utility customers. In such circumstances, although the customers choosing a CCA would not, arguably, directly pay their “full share” of the CRS cost components (CDWR purchase costs, bond related and administrative costs for the CDWR purchases, and other unrecovered costs), the benefits accruing to “others” from lower wholesale prices would more than make up for the under-recovered share of CRS costs, resulting in a win-win situation for all. Benefits to “others” does not end here. Some “new” customers would otherwise not move to California if required to pay a CRS set at the level paid by direct access customers. However, these new users might be attracted to California if the CPUC cut their full CRS burden and/or the CCA could reduce any CRS burden by reallocating its own aggregate CRS related cost recovery for its total load across its own retail customers. Simple arithmetic would show that any purely new customer load that pays any portion of the CRS would reduce the burden of all others. In brief, California is not a zero sum game. Finally, CCAs that build new efficient and/or renewable energy systems promote the state’s well-being, and improve employment in the state. This would provide additional direct benefits for all, including “others.” It is important to consider the last two categories: (1) truly incremental load that would at least wait to come to California until the CRS ends, and (2) incentives to build new renewables and create new jobs. These are incremental benefits that should be encouraged through public goods subsidies. Therefore, even CCAs that do not match their loads with new supplies should qualify for reduced CRS payments when they benefit “others.” In the long run, these cost savings for remaining utility customers will continue for the life of the plant. The CRS payments will expire as the costs associated with the CDWR contracts are paid. Thus, when the costs are paid down, remaining customers will receive an even greater benefit. These benefits will be lost if the CRS payments for departing customers, especially for new load that was never actually served by the CDWR contracts, prevent CCAs from building new cost-efficient, environmentally friendly generation plants in California. There is often a tendency for some to think that this is a zero sum game. Another simple example will help illustrate what we mean. Let’s say that

Handicapping the Winners

191

there was a cost of $1000 that could be recovered from 100 people. Some would aver that the only “fair” way to allocate to the cost would be for everyone to pay $10 (1000/100). These people would argue that it would be an inappropriate cost shift if some paid less than $10 while others paid more than $10. Similarly, if some paid $10 while others paid less than $10, these same folk would argue that not enough would be recovered to fully cover the cost. However, this kind of parochial thinking fails to consider the larger picture described above. Unequally allocating the costs makes sense if, by doing so, one creates benefits that either exceed the costs to be recovered or reduce the costs. By reducing the amount of the CRS to those CCAs that secure alternate sources of power or build their own generation, the price of electricity will be reduced, which could in turn help spur the state’s economy. In this situation, while unequally allocating the CRS costs might at first blush seem to under collect the costs or shift the costs to remaining IOU customers, the reverse is in fact true: prices and costs likely would be lowered to all the state’s residents, resulting in a win/win situation…not a zero sum game. There could be serious ramifications if new potential CCA customers, especially those that have never been provided generation services by an IOU, are forced to pay CRS charges that are set in such a way as to avoid any semblance of cost shifting. This would mean that California’s potentially new and highly competitive enterprise zones located at former military bases and elsewhere would be forced to pay an enormous tax to help bail out the IOUs and CDWR for money spent before these new commercial and industrial enterprises even open for business in California. Under such a scenario, it is likely that growth in California, especially growth in new enterprise zones, will be stymied. There is also another corollary to the price reduction benefits discussion above. Certain JPAs are dedicated to building state-of-the-art, environmentally friendly generating plants that combine renewables, such as solar, with natural gas combustion turbines. Such an approach helps attain the renewables standard mandated by the Legislature and provides an enormous public benefit. Imposing an unreasonably high CRS on CCA customers could very well foreclose building these innovative generating plants. There should be some CPUC approved targeted reduction in the CRS that corresponds to the benefits the public in general receives from building such fuel-efficient, environmentally friendly generating plants and/or that addresses localized unemployment. All Californians benefit from lower prices, cleaner air, greater fuel diversity, and economic growth. These beneficial activities should be rewarded by reducing the CRS to be paid by customers of a CCA that builds environmentally friendly generating plants.

192

The California Electricity Crisis: What, Why, and What’s Next

There is an additional issue with respect to whether the CRS overstates (and over-collects) certain costs. Recently, the Commission released about $1 billion the CDWR had held in reserve to back the CDWR purchases. This release resulted in a one-time refund to customers of about $40. The cost of these funds is included in the CRS to be collected in the future. Thus, in effect, the CRS is helping to fund this recent customer rebate. If the CRS is applied to “new” customers that were never served by the IOUs, this would mean that these customers are subsidizing the rate refund to “old” customers. This type of subsidy is clearly not justified. Creating different CRS levels raises complex issue of distinguishing between “old,” “new,” and “new-new” service. Due discrimination concepts are necessary to sort through these matters because charging everyone the same price is neither fair nor efficient. When CCAs secure new generation resources or build more generation, especially if that generation exceeds what is needed to serve their load, wholesale prices will be lowered for all Californians. This will help the state’s economy recover and cause even more MWHs to flow through the state’s grid, accelerating the new transition cost recovery and causing new infrastructure to be built that would help relieve congestion for all. Furthermore, two of the state’s newest seeds for energy recovery, JPAs and CCAs, would be encouraged to grow and complement one another. When “new” service produces benefits for “old” customers, these ripple effects should be considered in assessing the CRS level these “new” customers should pay. Just and reasonable, justified price differences are not necessarily foolish or unduly discriminatory. Pragmatic regulatory interpretations that recognize the role of due discrimination are important if the current economic malaise in California is to be resolved, and fairness, not necessarily equal CRS changes, is embraced.

CONCLUSION A pragmatic solution could set a somewhat arbitrary CRS for CCAs that secure alternate sources of generation or build new generation to serve their customers. An example might be as follows: CCA customers that cause new supply for California should pay a CRS that is 50 percent less than the base DA CRS established by the CPUC; An additional 25 percent reduction from the base DA CRS should be granted to those CCAs located in high unemployment areas; An additional 25 percent reduction from the base DA CRS for those CCAs utilizing renewable generation (over the state-mandated RPS level) to serve their customers’ load.

Handicapping the Winners

193

If at all possible, the CPUC would set these rules in advance so as to eliminate future debates and regulatory skirmishes. However, if no brightline test can be created, the Commission should establish a review process whereby CCA’s can demonstrate in their Implementation Plans (or annual reports) projects and activities justifying discounts from the otherwise applicable CRS. In such a way CCA projects and activities that benefit the direct participants and “others” will be encouraged, as a pragmatic “win/win” solution that benefits California and its citizens. California’s future would include municipalities and IOUs. The former would bring along IPPs. Competitive markets for long-term power and spot markets would emerge with no single dominant class of participants. At this point, no one can predict with any certainty what the future will bring. However, without a doubt, California is poised at a crossroads and must decide soon what market structure will be used to restore the state to its past glory.

This page intentionally left blank

Chapter 18 CONCLUSION: WRAPPING UP AND LESSONS LEARNED

California’s electricity crisis has hampered electricity industry restructuring in the state. Nevertheless, governments still seek gains in efficiency, new services, and lower prices. It is now clear that the vertically integrated utility is not a natural monopoly. More important, worldwide competition has much support. That said, California’s experience at least points to the complexity of change. More important, California’s experience should also strongly suggest that restructuring is not a one-step process. It is equally clear that markets need monitoring and streamlined rule-changing procedures when unintended or anomalous activity occurs or is detected. Through the mid 1970s, electricity prices generally declined, both in nominal and real terms. Simultaneously, sales grew, nearly doubling every decade. To accomplish this feat, utilities were required to continually seek technological improvements and to expand economies of scale. Given these pressures, it is not surprising that an engineering and technological focus dominated utility management. Customers were largely forgotten and marketing ignored. In hindsight, it is easy to see that technological progress would eventually begin to reach limits. This IOU engineering focus incorrectly presumed that growth would continue to reduce costs and that demand for electricity was inelastic and immune to competition. Most utilities were caught flatfooted when the world oil crisis hit and were trapped in “vicious cycle” of rising costs and falling demand. Energy utilities were attacked by politicians of all stripes. In most nations in the late 1970s, the electricity industry was the largest primary energy consumer of fossil fuels (oil, natural gas, and coal). The cost for these fuels was escalating rapidly. Electric utilities were also typically large private borrowers. The rapidly rising interest rates of this decade

196

The California Electricity Crisis: What, Why, and What’s Next

meant electric utilities had to pay significantly more to build power stations. The combination of rising fuel costs and rising interest rates caused electricity prices to increase. The local energy monopoly became a ready and easy scapegoat. Myopic and misdirected criticisms became the seeds for restructuring the electric utility industry and its regulatory institutions. The first break in the virtually integrated electric utility monopoly chain took place at the generation level when the Public Utility Regulatory Policy Act (1978) (PURPA) introduced a new class of independently owned and, therefore, alternative sources of generation known as Qualifying Facilities (QFs). These new QFs could be dispatched along with the traditional utilityowned generation. Regulators could count QFs as dependable sources of supply. Wholesale competition began. Independent power generators began to compete to attempt to take away large customers. Additionally, non-utility generators (NUGs) supplanted QFs and increased competition for large retail customers without becoming utilities. Various schemes were developed across the United States to entice retail load away from traditional IOU suppliers.1 The U.S. Congress enacted the Energy Policy Act (1992) to encourage greater transmission access to expand and encourage competitive wholesale power markets. This, in turn, encouraged additional requests for retail wheeling and direct access. The FERC responded by approving restructuring such as California’s 1996 efforts under AB 1890. The FERC has also moved from voluntary open-access transmission restructuring and a not-so-uniform Standard Market Design proposal for Wholesale Competition. The various restructuring models that were adopted in the United States to replace traditional monopoly-owned power station regulatory and ownership models (construction and ownership) with expanded wholesale competition, usually started in the wholesale electric power market with competitive bidding and purchase power agreements (PPA) for increments of new supply. Most restructuring in the United States remains focused on the wholesale power market. Before California, restructuring also seemed poised to expand into retail competition. It was somewhat surprising when in the spring of 1994, the CPUC announced plans for the most dramatic power sector restructuring ever attempted. The CPUC was prompted to act by high prices in California, a decade-and-a-half of experience under PURPA, and extensive experience reforming the telephone and natural gas industries. In retrospect, it is widely accepted that California’s restructuring was marked by a structurally flawed market design, was easily gamed, and had 1

For example, self-generation, joint NUG/industrial customer projects.

197 the unfortunate luck to experience a “perfect storm” of unexpected adverse market forces. These factors converged to cause the California Electricity Crisis from mid 2000 and to mid 2001. Our analysis reviewed what California attempted and why it went bad. Recent skeptics are claiming to have “told us so.” They seem to be fewer in number than the number of people who claimed to have designed California’s wonderfully complex system when wholesale generation prices fell to half of their cost of service of fully regulated levels during the first two years of restructured wholesale markets. We wrote this book to explore and learn from history. Our analysis demonstrates that California’s Electricity Crisis did not result simply from bad luck. Other factors were at work. There were structural design and regulatory flaws, primarily tied to an excessive nearly exclusive, reliance on spot markets. Worse, the regulatory reaction to crisis was slow and also flawed at times. California restructuring was based upon the concept that retail competition was both practical and inevitable. The CPUC wanted to quickly expand competition in California’s power markets to the retail level, while still protecting IOU shareholders and assuring lower retail prices. The CPUC approach introduced phrases such as “managed competition” and “phased deregulation.” These terms highlight the inherent conflict when governments try to have it both ways: simultaneously unleashing the competitive market and pulling on the political reins. California’s experience identifies some of the weakness inherent in mixed approaches. Managed competition or mixed-hybrid structures require monitoring, enforcement of rules, penalties for bad behavior, and streamlined processes for changing rules. Politicians and consumers also need to be prepared for price swings when the weather and market forces shift dramatically, as they did in 2000 in California and the west. More important, market participants need to hedge their positions against such unexpected shifts to insulate themselves from both economic and ex post political blame-game judgments. In theory, those who favor competition are not concerned with the vagaries and volatility inherent in competitive markets. Markets are uncertain and no one can predict with certainty what will happen in markets. This is why markets trump regulation. Economists fall into the market camp. They know that markets will have “winners and losers,” and accept the premise that market participants make choices and accept the consequences. Critics who prefer comprehensive results that are set and driven by politics, not competition, find this way of thinking to be reckless.

198

The California Electricity Crisis: What, Why, and What’s Next

The key difference is that economists and market restructuring architects need to build hedges into competitive wholesale markets. California’s fatal flaw was its nearly total reliance on spot markets, with little or no real opportunity to hedge ownership of generation or long term forward PPAs. It seems certain this mistake will not be made again. Two additional issues drive the political debate over introducing competition to electric power markets. First, utilities have invested billions of dollars under the existing regulatory scheme. These investments might be too expensive to compete effectively in a competitive market. In other words, the utilities would not be able to recover their investment and these assets would be stranded. However, retaining generation asset ownership is a built in hedge for utilities, as the California electricity crisis so aptly demonstrated. Therefore, new restructuring should focus less on stranded cost recovery and more on the value of ownership and the cost assignment for this reliability service. Second, protecting the environment and expanding DSM and “renewables” can be good for business. Reducing demand helps in times of crisis. California’s efforts here during the crisis are noteworthy. Under controlled competition or phased deregulation, the utility service customers (similar to core natural gas customers) who remain with the utility should be held harmless and should not pay for the departing (direct access) customers’ fixed cost recovery and associated return. Several tariff and accounting options can be used to protect remaining IOU customers. These include exit (or abandonment) fees, stand-by (or lean-on) fees, coordination (or reliability) fees, dispatch fees, transmission fees, distribution fees, and/or reinstatement (or prodigal child) fees. Some sort of “just and reasonable” fee for departing customers is also likely to be necessary. AB 117 shows that California’s retail consumers may be facing a second round of transition charges to recover the cost of the electricity crisis. The challenge is to be fair, while not discouraging innovation and competition. Fees, therefore, need to reflect benefits as well as costs. Restructuring opponents are lining up. They come in four varieties. First, there are shareholders and financial analysts. While they often do not trust regulators, they are especially wary when they observe IOU bankruptcy and ex post seller refunds in California. Second, economists and consultants support the theory of competition, but often give short shift to market monitoring and the ongoing need to refine the markets’ rules. Third, environmentalists often do not want to give up the political gains they have made with regulators that have caused IOUs to support both DSM and renewable investments. Fourth, retail consumers are wary that their restructuring may turn out to be as costly an experiment as was California wholesale restructuring, a cost they will have to pay.

199 California’s future is uncertain. There are five likely restructuring scenarios for the future. These are: (1) improve existing regulatory rules; (2) introduce wholesale power market competition by encouraging new entry of sellers and buyers; (3) introduce limited forms of retail competition; (4) expand retail competition; or, (5) full retail competition. In the first scenario, existing regulatory rules may be improved to specify an explicit utility performance contract and incentives for success. Rules could be changed to introduce performance-based regulation to improve cost of service regulations. Specific risks and responsibility would be assigned to customers, shareholders, and competitors. In the second scenario, wholesale power market competition would be expanded, increasing the nontraditional electricity supplier’s role in the wholesale power market. This scenario is the FERC’s current focus. This approach increases competitive bidding for new utility power requirements, interutility sales, and open access transmission. The third scenario introduces limited forms of retail competition. Utility investments (stranded costs and social programs) would need to be recovered. Competitive retail access would extend to utility transmission, with specific transmission access terms and pricing. The fourth scenario expands retail competition further by encouraging new retail brokers and aggregators and ESPs to replace or to compete against the utility as the exclusive electricity merchant, and redefines the utility’s duty to serve. The new municipal push in California is based on these ideas and the role of new buyers’ groups. Additionally, California is considering new non-core customer plans where larger customers would pay an exit fee and be allowed to purchase electricity in both the spot and long-term markets. The fifth scenario is the final stage: full retail competition with open access to dispatch centers, transmission, and information to enhance operational efficiency, with fully transparent transmission access and pricing protocols, neutral common carrier systems, and no duty to serve. This would, in effect, turn MWHs into internet-like service. This scenario now seems a remote prospect in the aftermath of California. The California crisis also points to the different roles played by Federal and State regulators. Many of these jurisdictional issues have not been sorted out in the United States. The legislative and judicial involvement in restructuring the United States also has not been fully determined. Forming a strategy or predicting the future in these circumstances is challenging. Nevertheless, the future is full of new opportunities. California likely will now go slow with electricity reforms. This latest California reform effort will likely focus on wholesale markets, not retail choice. The key is to have a significant number of new sellers and investors.

200

The California Electricity Crisis: What, Why, and What’s Next

This will not happen unless independent power producers return to California. This will not happen unless the CPUC creates a level playing field and eschews calls that favor IOU-owned generation. A second prong of this strategy is to have more buyers. This means releasing non-core customers from IOUs and encouraging cities that seek to form CCAs or new MOUs. Both require just, reasonable, and certain exit fee payments. Others must avoid California’s mistakes. This is too easily said to mean much. California restructured its electricity market knowing that there were unresolved issues and things that needed to be fixed. California’s spectacular early success caused too many to postpone these necessary tasks. When bad luck and the market conspired to push up prices, the political and regulatory climate were not suitable for making the needed reforms. Winners and losers began to play the blame game and sought advantages, not fundamental reforms. The primary lessons for others considering liberalized market reforms is to go slow, adopt a schedule for additional reforms, and keep to that schedule so that they are not forced to attempt to fix future problems after the problems become a crisis. Markets need monitoring, specific rules, and policing. Penalties need to be known before bad or anomalous behavior starts. Politicians and the public need to know at the outset that market prices will vary, both up and down. No one should claim credit for reductions and be allowed to blame others or the market when prices increase. Hedging through asset ownership and long-term contracts are the only market-oriented refuge for market participants that eschew price uncertainty and volatility. Regulation is still second best. California is now learning that re-regulation is not particularly easy, pretty, or wise. In California, what is certain is that investments in generation are needed sooner rather than later. Inertia is not an acceptable option if California is to avoid another crisis caused by supply shortages. A comprehensive political approach that begins by defining California’s future electricity structure is required. Such an approach will determine what entities will participate in this market. The legislature and new governor must act to set the agenda and establish a pragmatic and collaborative framework that will allow California to successfully emerge from this crisis and restore the state’s economy, economic growth, and state tax revenues. The CPUC must, in turn, expedite its decision making and reflect this comprehensive approach. The future is uncertain and will be challenging. But if the politicians, legislators, and regulators heed the lessons of the past, they will be able to

201 develop pragmatic solutions and approaches that will restore the luster to the Golden State.

This page intentionally left blank

Index

advertising 175 affiliate 177 antitrust 139 auction 61, 78 bankruptcy xvii, 17, 68, 159, 163, 174, 180, 198 baseload 12 benchmark 78, 79, 155, 156, 159, 160 bidding 45, 48, 63, 65, 78, 80, 129, 131, 133, 141, 142, 160, 196, 199 bilateral 66, 68, 137, 150, 155 Borenstein 79 brokers 49, 199 Cicchetti 207 climate 1, 3, 72, 90, 94, 110, 135, 170, 175, 187 collusion 142 competition xvii, 6, 16, 17, 18, 25, 26, 35, 55, 73, 89, 141, 157, 175, 184, 195, 196, 197, 198, 199 concentration 52 congestion 10, 48, 55, 64, 129, 130, 132, 133, 140, 144, 192 conservation 53, 54, 173, 183, 188 conspiracy 2, 138 credit 68, 72, 130, 150, 161, 163 creditworthiness 68, 159

daily 69, 72, 85, 86, 88, 90, 91, 92, 93, 94, 95, 99, 101, 103, 104, 105, 107, 111, 113, 117, 119, 164, 165, 166 damage xvii debate 32, 138, 174, 198 demand 6, 7, 8, 9, 10, 11, 12, 13, 17, 23, 26, 27, 28, 32, 33, 34, 35, 36, 38, 41, 42, 43, 44, 52, 55, 56, 57, 58, 59, 60, 62, 64, 65, 67, 69, 70, 71, 72, 73, 78, 79, 80, 86, 87, 88, 89, 90, 91, 94, 101, 104, 108, 110, 119, 127, 128, 129, 133, 137, 138, 143, 147, 148, 149, 151, 153, 154, 155, 168, 169, 170, 171, 173, 174, 179, 182, 189, 195, 198 depreciation 19, 20, 21 deregulate 23 deregulation 23, 57, 155, 174, 182, 197, 198 discrete 76, 108 discriminatory 6, 18, 187, 192 diseconomies 17 disinvestment 32 Dubin 108, 208 econometric xviii, 3, 71, 75, 77, 79, 81, 83, 90, 94, 107, 125 efficiency 6, 9, 17, 22, 23, 26, 27, 28, 32, 34, 54, 139, 140, 141, 157, 173, 182, 183, 188, 195, 199 elasticity 23, 35, 36, 38, 41, 44, 79, 94, 95

204 emergency 65, 73, 105, 108, 110, 113, 124, 130, 145, 160, 161, 162 endogeneity 124 engineering 15, 78, 79, 195 Enron 128, 130, 131, 132, 134, 135, 143, 144, 145, 152, 163, 165, 208, 209 enterprise 6, 176, 177, 178, 191 entrepreneurs 1 entry 16, 18, 44, 45, 178, 182 environment 198 Europe 103 exogenous 108 exorbitant 54, 65 experiment 145, 198 facilities 173, 182 facility 141, 181 factor 10, 12, 56, 57, 94, 108, 1 1 1 , 113, 124, 169, 185, 189 failure 3, 54, 57, 65, 73 falsifying 132 finance 77, 176 forecast 11, 23, 65, 77, 171 foreclose 191 forward 27, 66, 67, 68, 73, 103, 149, 150, 155, 156, 198 franchise 6, 15, 32, 147, 176, 178 games 138, 144, 145, 187 gaming 3, 128, 130, 134, 137, 138, 139, 140, 141, 143, 144, 145, 147, 161, 162, 174 gasoline 94 generation 2, 9, 15, 22, 23, 25, 27, 33, 45, 47, 48, 52, 53, 54, 55, 56, 57, 58, 60, 62, 63, 64, 67, 70, 80, 88, 90, 106, 107, 108, 133, 137, 138, 141, 155, 156, 163, 173, 174, 177, 178, 179, 180, 181, 182, 183, 185, 188, 189, 190, 191, 192, 196, 197, 198 generator 69, 88, 151, 152, 159, 161, 181 Harvey 81 Hatanaka 1 1 1 Hausman 124 hedge 67, 137, 141, 197, 198 hedging 27, 63, 66, 68, 103, 127, 130, 139 Herfindahl 86

Index Hirschman 86 Hogan 81 Houthhakker 94 hydroelectric 54, 56, 58, 60, 87, 106, 108, 111, 168 hydropower 54, 141 identification 107, 111 incumbent 23, 52, 55, 157, 176, 178 indebtedness 173 industrial 17, 20, 21, 22, 23, 175, 191, 196 inefficient 11, 138, 141 inelastic 27, 41, 42, 195 inelasticity 108 institutional 2, 71, 72, 125, 169, 171 instrumental 107 instruments 66, 107 ISO 48, 69, 105, 139, 140, 147, 160, 163 Joskow xvii, 79 Kahn 16, 17, 18, 19, 79 Koyck 76 lambda 9, 32 liability 134, 145, 161, 163, 164, 175, 185,186 load 7, 8, 9, 10, 11, 12, 13, 15, 32, 33, 64, 77, 105, 108, 113, 114, 119, 128, 129, 143, 155, 161, 185, 186, 188, 189, 190, 192, 196 marginal 3, 9, 10, 12, 13, 22, 25, 26, 27, 29, 31, 32, 33, 34, 35, 38, 39, 40, 41, 42, 43, 44, 45, 56, 60, 61, 62, 72, 79, 80, 83, 87, 107, 142, 153, 159, 160, 161, 163, 187 market xvii, 2, 3, 15, 17, 25, 26, 27, 28, 30, 31, 32, 33, 44, 45, 47, 48, 49, 51, 53, 54, 55, 56, 59, 60, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 75, 77, 78, 79, 80, 83, 85, 86, 87, 89, 90, 91, 93, 95, 101, 104, 105, 106, 108, 111, 113, 117, 119, 124, 125, 127, 128, 129, 130, 131, 132, 133, 134, 135, 137, 138, 139, 140, 141, 142, 143, 144, 145, 147, 148, 149, 150, 151, 152, 153, 154, 155, 156, 157, 159, 160,

Index 161, 162, 163, 164, 165, 168, 169, 170, 171, 174, 178, 188, 189, 193, 196, 197, 198, 199 markup 35, 42 McCullough 83 MCF 60 McFadden 108 MCP 25, 26, 49, 55, 60, 61, 64, 80, 83, 160, 161, 163, 164 megawatt 7, 61, 152, 155 Merton 75 mitigation 160, 161, 162, 163, 165 MMCP 78, 80, 162, 163, 164, 165, 166 monopolist 15, 41, 42 monopoly 5, 6, 15, 16, 17, 18, 23, 25, 33, 35, 39, 41, 138, 142, 157, 195, 196 multicollinearity 72, 73, 77 multiplier 38 municipal 69, 131, 175, 177, 181, 183, 185, 186, 199 municipalities 174, 175, 176, 177, 178, 181, 182 Nerlove 76 network 10, 47, 48 NOAA 105 optimize 58 optimizing 173 outage 63 outflow 144 overscheduling 129, 138, 143 overseers 138 Pacific xvii, 56, 87, 128 Papoulis 76 payment 6, 153, 178, 189 peaker 11 penalties 135, 153, 159, 197 penalty 134, 154, 155 permit 5 petroleum 72, 90, 127, 170 polity 18 pollution 53, 138, 188 portfolio 20, 164, 183 prediction 101, 107 price 3, 5, 6, 16, 19, 22, 23, 25, 26, 27, 30, 31, 32, 33, 34, 35, 36, 38, 39, 40, 41, 42, 44, 48, 51, 55, 56, 57, 58, 60,

205 61, 62, 64, 65, 66, 67, 69, 70, 71, 72, 73, 76, 77, 78, 79, 83, 85, 89, 90, 91, 92, 93, 94, 95, 98, 101, 104, 106, 107, 108, 113, 119,124, 125, 130, 131, 137, 141, 142, 143, 145, 148, 149, 150, 151, 153, 154, 155, 160, 161, 162, 163, 164, 165, 166, 168, 169, 170, 171, 177, 182, 186, 187, 189, 191, 192, 197 probability 7, 8, 12, 33, 94, 108, 111 producer 26, 34, 44 production 2, 9, 15, 27, 29, 30, 33, 60, 62, 80, 85, 86, 87, 142, 152, 159 productivity 9 punish 129, 174

rain 105, 111 random 76, 77 rate xvii, 9, 17, 18, 19, 21, 23, 30, 44, 56, 105, 108, 113, 152, 156, 161, 163, 164, 180, 181, 192 ratepayers 17, 144 reclaim 62, 72 redispatch 64 refund 80, 101, 105, 108, 113, 114, 119, 124, 125, 134, 142, 145, 152, 160, 161, 162, 163, 164, 165, 166, 192 regional 56, 59, 60, 89, 104, 107, 108, 113, 131, 161, 176, 180 regions 54, 117, 173 regression 19, 77, 91, 93, 94, 97, 98, 99, 107, 108, 111, 112, 119, 124 regulation xvii, 2, 3, 5, 13, 15, 16, 17, 18, 22, 23, 25, 31, 35, 39, 40, 41, 42, 44, 45, 53, 54, 64, 66, 139, 145, 157, 177, 181, 184, 197, 199 regulator 19 regulatory 3, 13, 15, 16, 17, 18, 19, 20, 22, 23, 27, 32, 36, 38, 39, 41, 42, 44, 54, 55, 75, 100, 101, 104, 105, 108, 113, 119, 124, 138, 141, 143, 144, 145, 156, 157, 160, 162, 163, 164, 177, 181, 182, 187, 192, 193, 196, 197, 198, 199, 209 rehearing 187 reliability 47, 48, 55, 63, 64, 65, 69, 104, 106, 169, 174, 182, 198 rent 19, 80 restructure 22, 23, 45, 47, 54, 153, 180

206 rule 9, 33, 35, 38, 41, 44, 154, 157, 164, 195 rulemaking 32 scarcity 108, 152, 153, 154, 155 scenario 191, 199 season 7 seller 128, 131, 132, 138, 142, 152, 154, 155, 183, 198 serial 81, 92, 107, 111 shareholder 135 shortfall 89 signal 45, 108, 169 simultaneous 130 stakeholder 129, 143, 144, 145, 159 statistical 3, 85, 90, 91, 92, 95, 104, 107, 111, 113, 114, 119, 124, 171 statute 177 stochastic 76, 77 stranded 55, 156, 198, 199 subsidize 17, 54, 183 subsidy 192 supply 5, 6, 9, 12, 15, 18, 21, 26, 27, 28, 31, 33, 42, 44, 52, 54, 55, 56, 57, 58, 59, 60, 62, 64, 65, 66, 70, 78, 79, 80, 86, 87, 88, 89, 90, 91, 94, 98, 104, 107, 108, 110, 113, 120, 127, 128, 133, 137, 138, 142, 143, 149, 151, 153, 154, 155, 157, 168, 169, 170, 173, 174, 181, 188, 189, 192, 196 surcharge 60, 185 surveillance 140, 144, 148, 153 synergy 32 tariffs 21, 22, 23, 25, 33, 44, 45, 55, 130, 140, 144, 145, 147 tax 21, 176, 177, 191 temperature 77, 105 trader 131 transition 6, 55, 192, 198 transmission 10, 15, 48, 54, 55, 63, 64, 130, 132, 140, 173, 178, 182, 183, 186, 188, 196, 198, 199

Index turbine 9 Turvey xviii, 38, 39, 44 unbundled 5 uncertainty 137, 184 underschedule 65 unemployment 105, 108, 113, 119, 188, 191, 192 utilities 1, 5, 6, 12, 16, 17, 18, 19, 32, 42, 44, 45, 52, 53, 54, 55, 60, 67, 68, 69, 87, 131, 149, 150, 175, 177, 178, 182, 183, 187, 195, 196, 198 utility 5, 6, 9, 15, 16, 17, 18, 19, 20, 21, 22, 23, 36, 38, 44, 45, 48, 68, 69, 106, 129, 147, 150, 156, 161, 173, 175, 177, 181, 183, 185, 186, 187, 188, 190, 195, 196, 198, 199, 208 variables 72, 77, 91, 92, 93, 95, 98, 104, 106, 107, 108, 111, 113, 117, 119, 124 variance 76, 112, 113, 114, 119 Verlag 76 Verleger 94 volatile 27, 66, 90, 127, 170 volatility 27, 65, 76, 103, 174, 197 Wagner 91 Weiner 76 Weintraub 35 welfare 34, 36, 38, 39, 41, 42 wheeling 196 Williams 95 Williamson 42 withhold 138, 139 withholding 141, 153, 155 Wolak 78, 79, 103 Wolfram 78, 79 yield 19, 35, 38, 161, 184 zonal 48

About the Authors

CHARLES J. CICCHETTI Dr. Charles J. Cicchetti holds the Miller Chair in Government, Business, and the Economy at the University of Southern California. He is the former Managing Director of Arthur Andersen Economic Consulting and former Co-Chairman of Putnam, Hayes & Bartlett, Inc., and Deputy Director of the Energy and Environmental Policy Center at Harvard University’s John F. Kennedy School of Government. He was a co-founder of Madison Consulting Group, which merged with National Economic Research Associates, Inc. (NERA), where he served as Senior Vice President. Dr. Cicchetti is a former regulator who chaired the Wisconsin Public Service Commission and directed the Wisconsin Energy Office. He has worked in developing nations throughout the world and has served on numerous federal and international energy and environmental committees. The author of numerous books and articles, Dr. Cicchetti’s publications include Alaskan Oil: Alternative Routes and Markets; Perspectives on Power, co-authored with Edward Berlin and William Gillen; The Marginal Cost and Pricing of Electricity: An Applied Approach, with William Gillen and Paul Smolensky; The Costs of Congestion: An Economic Analysis of Wilderness Recreation, with V. Kerry Smith; and Forecasting Recreation in the United States. He has edited Energy Systems Forecasting, Planning, and Pricing, with W.K. Foell; and Studies in Electric Utility Regulation, with John Jurewitz. He has recently completed an update to Restructuring Electricity Markets: A World Perspective entitled Restructuring Electricity

208

About the Authors

Markets: A World Perspective Post-California and Enron. Dr. Cicchetti attended the United States Air Force Academy, received a BA from Colorado College, and a Ph.D. from Rutgers University, both in economics. He did post-doctoral research at Resources for the Future, served as chief economist for the Environmental Defense Fund, and was a professor of economics and environmental studies at the University of Wisconsin, Madison. He is currently teaching environmental and energy economics at the University of Southern California in Los Angeles.

JEFFREY A. DUBIN Dr. Jeffrey A. Dubin is co-founding member and Director of Statistical and Economic Analysis at Pacific Economics Group. He is currently an Associate Professor at the California Institute of Technology, where he has been a faculty member since 1982. Dr. Dubin earned his undergraduate degree in Economics with highest honors and great distinction from the University of California, Berkeley, and received a Ph.D. in Economics from the Massachusetts Institute of Technology. Dr. Dubin’s research focuses on microeconomic modeling with particular emphasis in applied econometrics. His current research concerns contingent valuation methods, discrete choice econometrics, effects of welfare and entitlement programs on unemployment, energy economics and tax compliance. He is the author of numerous publications, including Consumer Durable Choice and the Demand for Electricity, Elsevier North-Holland, 1985; Studies in Consumer Demand—Econometric Methods Applied to Market Data, Kluwer Academic Publishers, 1998; and most recently, Empirical Studies in Applied Economics, Kluwer Academic Publishers, 2001. He received the Econometric Society Frisch Medal in 1986 with Professor Daniel McFadden.

COLIN M. LONG Colin M. Long has more than twenty years’ litigation and consulting experience, specializing in energy and utility-related matters. He is an experienced federal and state court litigator at both the trial and appellate level. He has extensive experience working with regulated industries,

About the Authors

209

particularly in cases involving mergers and acquisitions and matters before the FERC and state regulatory commissions. Mr. Long received a BA in political science, cum laude, from the University of Southern California and a J.D. from Loyola University School of Law. He is admitted to practice in the State of California, the Ninth Circuit District Courts, the Ninth Circuit Court of Appeals, and the United States Supreme Court. He is a member of the American Bar Association and the Association of Business Trial Lawyers. Mr. Long has co-authored several recent articles with respect to energy issues, including “ISOs and Transcos: What’s At Stake,” “Politics As Usual: A Roadmap to Backlash, Backtracking, and Re-Regulation,” and “Transmission Products and Pricing: Hidden Agendas,” and with Dr. Cicchetti recently completed a book entitled Restructuring Electricity Markets: A World Perspective Post-California and Enron.

The California Electricity Crisis (HOOVER INST PRESS PUBLICATION)

ESTONIA AND ESTONIANS (HOOVER INST PRESS PUBLICATION)

AZERBAIJANI TURKS (HOOVER INST PRESS PUBLICATION)

The Road Ahead for the Fed (HOOVER INST PRESS PUBLICATION)

Liberty And Justice (Hoover Institution Press Publication)

Liberty and Democracy (Hoover Institution Press Publication)

Teacher Quality (Hoover Institution Press Publication)

Ending Government Bailouts as We Know Them (HOOVER INST PRESS PUBLICATION)

Communist Regimes in Eastern Europe (Hoover Press publication)

The California Electricity Crisis: What, Why, And What's Next

California's Electricity Crisis

Modern Uzbeks: From the 14th Century to the Present : A Cultural History (Hoover Institution Press Publication)

Do the Right Thing: The People's Economist Speaks (Hoover Institution Press Publication)

Power To The Patient: Selected Health Care Issues And Policy Solutions (Hoover Institution Press Publication)

Russia's 1996 Presidential Election: The End of Polarized Politics (Hoover Institution Press Publication)

Our Brave New World: Essays on the Impact of September 11 (Hoover Institution Press Publication)

Competing With the Government: Anticompetitive Behavior and Public Enterprises (Hoover Institution Press Publication)

Liberty and Research and Development: Science Funding in a Free Society (Hoover Institution Press Publication)

The Hoover Dam

Property Rights: A Practical Guide to Freedom and Prosperity (Hoover Institution Press Publication)

ELECTRICITY

Electricity

Liberal Reform in an Illiberal Regime: The Creation of Private Property in Russia, 1906-1915 (Hoover Institution Press Publication)

The University of California Press: the early years, 1893-1953

California

California

California

The story of electricity

Electricity Demystified

Electricity Demystified

Hoover Dam: An American Adventure

The California Electricity Crisis (HOOVER INST PRESS PUBLICATION)

ESTONIA AND ESTONIANS (HOOVER INST PRESS PUBLICATION)

AZERBAIJANI TURKS (HOOVER INST PRESS PUBLICATION)

The Road Ahead for the Fed (HOOVER INST PRESS PUBLICATION)

Liberty And Justice (Hoover Institution Press Publication)

Liberty and Democracy (Hoover Institution Press Publication)

Teacher Quality (Hoover Institution Press Publication)

Ending Government Bailouts as We Know Them (HOOVER INST PRESS PUBLICATION)

Communist Regimes in Eastern Europe (Hoover Press publication)

The California Electricity Crisis: What, Why, And What's Next

California's Electricity Crisis

Modern Uzbeks: From the 14th Century to the Present : A Cultural History (Hoover Institution Press Publication)

Do the Right Thing: The People's Economist Speaks (Hoover Institution Press Publication)

Power To The Patient: Selected Health Care Issues And Policy Solutions (Hoover Institution Press Publication)

Russia's 1996 Presidential Election: The End of Polarized Politics (Hoover Institution Press Publication)

Our Brave New World: Essays on the Impact of September 11 (Hoover Institution Press Publication)

Competing With the Government: Anticompetitive Behavior and Public Enterprises (Hoover Institution Press Publication)

Liberty and Research and Development: Science Funding in a Free Society (Hoover Institution Press Publication)

The Hoover Dam

Property Rights: A Practical Guide to Freedom and Prosperity (Hoover Institution Press Publication)

ELECTRICITY

Electricity

Liberal Reform in an Illiberal Regime: The Creation of Private Property in Russia, 1906-1915 (Hoover Institution Press Publication)

The University of California Press: the early years, 1893-1953

California

California

California

The story of electricity

Electricity Demystified

Electricity Demystified

Hoover Dam: An American Adventure

Recommend Documents